Background
Sudden death is one of the characteristics of Chagas disease (ChD). With the development of strategies for the prevention of malignant arrhythmias, especially with implantable cardioverter-defibrillators (ICDs), there is interest in developing strategies to predict sudden cardiac death. The aim of this study was to test the hypothesis that global longitudinal strain (GLS) and mechanical dispersion (MD) may be associated with malignant ventricular arrhythmias in patients with ChD.
Methods
A cross-sectional study was conducted including 62 patients with ChD who were separated into two groups according to ICD implantation status. Group 0 consisted of 34 patients with ChD without ICDs, and group 1 comprised 28 patients with ICDs. Complete echocardiographic studies, including GLS and MD measurements, were performed in all patients.
Results
Chamber dimensions, ejection fraction, and diastolic function showed no significant differences between patients with and those without ICDs. GLS was reduced in patients with ChD with ICDs compared with those without ( P = .02). By receiver operating characteristic curve analyses, GLS identified patients with ChD with ICDs with sensitivity of 67% and specificity of 69%. MD was more pronounced in patients with ChD with ICDs compared with those without ( P < .001), with a C statistic of 0.83 (95% CI, 0.71–0.91). MD > 57 msec detected ICD presence with sensitivity of 79% and specificity of 71% and was superior to GLS and ejection fraction ( P < .05). In multivariate analysis, New York Heart Association functional class (odds ratio, 3.02; 95% CI, 1.09–8.39; P = .03), MD (odds ratio, 1.11; 95% CI, 1.04–1.19; P = .001), and GLS (odds ratio, 0.72; 95% CI, 0.54–0.96; P = .026) were significant and independently associated with malignant arrhythmic events.
Conclusions
GLS and MD may add important information in the risk stratification of patients with ChD. The use of MD by strain echocardiography could be an attractive tool in the decision making for ICD placement as primary prevention for sudden cardiac death in patients with ChD.
Chagas disease (ChD) remains a leading medicosocial problem in Latin America, affecting 8 million to 10 million people, with high morbidity and mortality and a severe economic impact. Although many individuals remain asymptomatic for long periods, approximately one third of infected patients develop life-threatening heart disease, including malignant ventricular arrhythmias and heart failure.
Sudden death has been one of the characteristics of ChD since the first descriptions of the disease and is responsible for >50% of mortality in patients with ChD with heart failure. The electrophysiologic mechanisms most frequently involved in sudden cardiac deaths in ChD are ventricular tachycardia (VT) and ventricular fibrillation (VF). With the development of therapeutic strategies for the prevention of death due to VF, especially with implantable cardioverter-defibrillators (ICDs), there is considerable interest in developing strategies to predict disease progression and arrhythmias in patients with ChD.
ICD implantation as primary prevention in patients with ChD remains a controversial matter. Although several clinical features have been assessed as possible risk factors for the development of fatal arrhythmias, the sensitivity and specificity of any single test to predict fatal arrhythmias has been limited.
Myocardial strain by echocardiography can accurately quantify regional myocardial timing and function. Global longitudinal strain (GLS) has been shown to be more accurate than left ventricular (LV) ejection fraction (LVEF) in quantifying LV function and also to possess prognostic impact. Myocardial mechanical dispersion by strain echocardiography is a sensitive measure of inhomogeneous ventricular contractions, and recent studies have shown that mechanical dispersion predicted ventricular arrhythmias both in patients with cardiomyopathies and after myocardial infarctions. In the present study, we hypothesized that GLS and mechanical dispersion may be associated with malignant arrhythmias in patients with ChD and therefore may contribute to risk stratification of arrhythmias in these patients.
Methods
From January 2012 through July 2014, we conducted a cross-sectional study including 62 patients with ChD selected from the Referral Center for Training on Infectious and Parasitic Diseases of the Federal University of Minas Gerais. Included patients were split into two groups according to ICD implantation status. Group 0 consisted of patients with Chagas dilated cardiomyopathy characterized by LV diastolic diameter indexed to body surface area ≥ 31 mm/m 2 and LVEF < 55%, representing the most severe form of ’ChD, but with no previous malignant ventricular arrhythmias and no ICDs. Group 1 comprised patients with ChD with ICDs implanted as secondary prevention, according to the guidelines for ICD implantation defined by the Brazilian Ministry of Health :
- •
Resuscitated from cardiac arrest (documented sustained VT or VF) due to nonreversible pathology, with LVEF ≤ 35% or structural heart disease
- •
Spontaneous VT due to nonreversible causes, with LVEF ≤ 35% or structural heart disease
- •
Syncope of unknown origin, with inducibility of hemodynamically unstable or clinically relevant VT or VF in electrophysiologic study, with LVEF ≤ 35% or structural heart disease
Therefore, all patients with ICDs had experienced events of ventricular arrhythmia or were inducible during electrophysiologic studies after cardiac syncope. ICD implantation was used as a surrogate for “malignant arrhythmias” in group 1. None of the patients had ventricular pacing during the echocardiographic examination.
All patients underwent clinical examination and resting electrocardiography. The QT interval was measured by the tangent method and corrected for heart rate according to Bazett’s formula. Left bundle branch block was defined as QRS duration ≥ 120 msec; broad notched or slurred R waves in leads I, aVL, V 5 , and V 6 ; absent q waves in leads I, V 5 , and V 6 ; R peak time > 60 msec in leads V 5 and V 6 ; and ST and T waves usually opposite in direction to QRS. Right bundle branch block was defined as QRS duration ≥ 120 msec; rsr′, rsR′, or rSR′ in lead V 1 or V 2 ; S wave of greater duration than R wave or >40 msec in leads I and V 6 ; and normal R peak time in leads V 5 and V 6 but >50 msec in lead V 1 .
The research project was evaluated and approved by the Committee of Research Ethics of the Federal University of Minas Gerais. ChD was defined as positive serum tests for Trypanosoma cruzi in at least two of the three different techniques available (reaction to indirect immunofluorescence, indirect hemagglutination, and enzyme-linked immunosorbent assay).
Patients with the following comorbidities were excluded: systemic arterial hypertension, coronary artery disease, rheumatic diseases, diabetes mellitus or glucose intolerance, thyroid dysfunction, kidney failure, chronic obstructive pulmonary disease, hydroelectrolytic disorders, significant anemia, and pregnancy.
Echocardiography
The echocardiographic studies were performed using S6 and E9 systems (GE Vingmed Ultrasound AS, Horten, Norway) and analyzed using commercially available software (EchoPAC; GE Healthcare, Milwaukee, WI). LVEF was assessed using Simpson’s biplane method. By speckle-tracking echocardiography, longitudinal strain was obtained from apical four-chamber, two-chamber, and long-axis views. Three cardiac cycles from each view were recorded for offline analyses with a frame rate > 50 frames/sec. Peak negative longitudinal strain was assessed in 16 LV segments, defined as the peak negative value during the entire cardiac cycle, hence including postsystolic shortening, and was averaged as GLS. The time interval from electrocardiographic onset of Q/R to peak negative strain was assessed in each of the 16 LV segments ( Figure 1 ). Mechanical dispersion was defined as the SD of the time to peak negative strain in the same 16 LV segments. Bull’s-eye plots showing regional strain amplitudes were constructed for patients with ChD with and without ICDs. Mitral inflow E velocity was recorded using pulsed Doppler. The e′ velocity (tissue Doppler) was averaged from the septal and lateral mitral annuli, and the E/e′ ratio was calculated. In group 1, echocardiographic examinations were performed after ICD placement. All echocardiographic studies and analyses were performed by the first author (M.V.L.B.).
Statistical Analysis
Clinical and echocardiographic data were analyzed using SPSS version 18 (SPSS, Chicago, IL). Categorical variables were compared using χ 2 tests; continuous variables, expressed as mean ± SD, were compared using unpaired Student’s t tests. Logistic regression analyses were performed to identify predictors for ICD implantation. Odds ratios and 95% CIs were calculated. The multivariate analysis was performed by including significant variables from the univariate model ( P < .05). For LVEF, GLS, and mechanical dispersion, receiver operating characteristic curves were constructed. Reproducibility of GLS was expressed as intraclass correlation coefficient for repeated measures in a random sample of 15 patients. P values < .05 were considered to indicate statistical significance.
Results
Study Population
We included 62 patients with ChD (mean age, 58.3 ± 8.3 years; 62% male). Of these, 28 had ICD implantation due to previous arrhythmic events, as defined in the “Methods” section, whereas 34 were included with no previous arrhythmic events, hence not fulfilling criteria for secondary prevention ICD placement. Clinical characteristics of the study patients stratified by the presence of ventricular arrhythmia are summarized in Table 1 . Twenty-three patients (37%) were in New York Heart Association functional class I at the time of recruitment into the study. The patients who presented malignant arrhythmic events were older and more symptomatic than those patients without previous arrhythmic events.
| Clinical data | Group 0 ( n = 34) | Group 1 ( n = 28) | P |
|---|---|---|---|
| Age (y) | 55.8 ± 8.3 | 60.5 ± 7.8 | .026 |
| Male gender | 21 (62%) | 18 (64%) | .838 |
| NYHA functional classes II–IV | 18 (53%) | 21 (75%) | .036 |
| RV failure ∗ | 8 (23%) | 7 (25%) | .650 |
| Heart rate (beats/min) | 66.1 ± 9.3 | 62.8 ± 7.3 | .258 |
| Systolic blood pressure (mm Hg) | 112.1 ± 17.2 | 116.4 ± 21.3 | .531 |
| Diastolic blood pressure (mm Hg) | 71.9 ± 11.9 | 76.6 ± 12.1 | .282 |
| Medications | |||
| Diuretics | 20 (59%) | 15 (54%) | .771 |
| ACE inhibitors | 21 (62%) | 6 (21%) | .005 |
| β-blockers | 24 (70%) | 23 (82%) | .433 |
| Anticoagulants | 6 (18%) | 3 (11%) | .341 |
| Amiodarone | 13 (38% | 15 (54%) | .288 |
Electrocardiographic Findings
QRS duration was increased in all patients with ChD, but no difference was observed between those with and those without ICD. Corrected QT interval duration was within normal limits and was similar in the two groups ( Table 2 ). No differences were found in the prevalence of intraventricular (right or left) branch block between the groups ( P > .05).
| Parameter | Group 0 ( n = 34) | Group 1 ( n = 28) | P |
|---|---|---|---|
| QRS duration (msec) | 129 ± 27 | 122 ± 30 | .54 |
| Corrected QT interval (msec) | 451 ± 36 | 453 ± 38 | .88 |
| RBBB | 11(32%) | 8 (29%) | .79 |
| LBBB | 6 (18%) | 4 (14%) | .68 |
| LVEDD (mm) | 65 ± 8 | 64 ± 7 | .59 |
| LVESD (mm) | 52 ± 10 | 51 ± 10 | .52 |
| LAD (mm) | 44 ± 6 | 45 ± 5 | .74 |
| LVEF (%) | 41 ± 14 | 39 ± 11 | .52 |
| E/e′ ratio | 10.4 ± 4 | 10.0 ± 3 | .69 |
| Mechanical dispersion (msec) | 49 ± 21 | 82 ± 29 | <.001 |
| GLS (%) | −16.5 ± 4.3 | −13.6 ± 5.5 | .022 |
Echocardiographic Findings
Chamber dimensions were severely enlarged, LVEF was reduced, and diastolic function was impaired in both groups of patients with ChD compared with established reference values. No differences in these parameters were observed between those with and those without ICDs ( Table 2 ). However, systolic function by GLS was worse in patients with ChD with arrhythmias compared with those without arrhythmias ( P = .02) ( Table 2 ). Segmental LV wall motion analyses by visual assessment showed abnormalities most frequently in the inferior (28.1%), inferolateral (37.5%), and inferoseptal (20.3%) walls. Apical aneurysms were found in 27.6% of the patients. Strain analyses showed that regional dysfunction was located mainly in the inferoseptal and inferolateral walls in patients with ChD ( Figure 2 ). In patients with ChD with arrhythmias, function in the antero- and inferoseptal segments was worse compared with patients with ChD without arrhythmias ( P < .05) ( Figure 2 ).
By receiver operating characteristic curve analyses, GLS showed fair discriminative ability between patients with and those without ICDs, with a C statistic of 0.65 (95% CI, 0.49–0.79) ( Figure 3 ). The optimal cutoff value for GLS was −14.3%, with sensitivity of 67% and specificity of 69%.
Mechanical dispersion showed the best discrimination between the two groups, with a C statistic of 0.83 (95% CI, 0.71–0.91), being superior to both GLS ( P < .001) and LVEF ( P < .001). Mechanical dispersion of 57 msec was the optimal cutoff value and detected patients with ICDs and hence arrhythmic events with sensitivity of 79% and specificity of 73% ( Figure 3 ). In multivariate analysis, both GLS and mechanical dispersion were significant and independent markers of history of malignant ventricular arrhythmias (group 1, patients with ICDs) ( Table 3 ), after adjustment for New York Heart Association functional class.
