Author + information
- Institut Clínic Cardio-Vascular (ICCV), Hospital Clínic, Universitat de Barcelona, Catalonia, Spain; Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona, Spain; and the Centro Nacional de Investigaciones Cardiovasculares Carlos III (CNIC), Madrid, Spain
- ↵∗Address for correspondence:
Dr. Ana García-Álvarez, Cardiology Department. Clinic Cardiovascular Institute, Hospital Clínic, Villarroel 170, 08027, Barcelona, Spain.
Chagas disease remains a major cause of morbidity and mortality in several countries of Latin America, where the disease is endemic, and a growing worldwide problem as a result of migration flows. Chagas heart disease (ChHD) constitutes the most serious form of the chronic phase of the disease and represents the main cause of mortality in these patients. Several variables have been shown to predict mortality in ChHD, with left ventricular ejection fraction (LVEF) the most powerful independent predictor (1). In 2006, Rassi et al. (2) published a prognostic score for global mortality derived from a retrospective follow-up (mean: 7.9 years) of 424 patients with ChHD and validated in 153 patients. This score, which revealed a C statistic of 0.84 in the development cohort and 0.81 in the validation cohort, included New York Heart Association (NYHA) functional class III or IV, cardiomegaly, left ventricular (LV) systolic dysfunction on echocardiography, nonsustained ventricular tachycardia (VT) on 24-h Holter monitoring, low QRS voltage on electrocardiography, and male sex.
Persistent myocardial inflammation and progressive fibrosis represent the main pathological characteristics at histology. In addition, the incidence of ventricular arrhythmias, the leading cause of sudden cardiac death, is associated with the progression and extension of myocardial disease, and, probably is especially related to re-entry pathways generated by the patchy fibrosis. This myocardial fibrosis can be detected by cardiac magnetic resonance (CMR) using late gadolinium enhancement (LGE) sequences, which have been shown to be an independent prognostic marker in several other cardiomyopathies (3,4).
In patients with ChHD, several previous studies have shown a significant association between the presence of LGE and LV dysfunction (5,6) and sustained VT (6–8). In a study by Rochitte et al. (6), all patients with previous documented VT (n = 10) had extensive LGE (25.4 ± 9.8%). Mello et al. (7) included 41 patients with ChHD with positive LGE and segmental wall motion abnormalities and classified them according to history of sustained VT. There were no statistical differences regarding the magnitude of LGE (VT group, 30.0 ± 16.2%; non-VT group, 21.7 ± 15.7%; p = 0.118), but the presence of ≥2 contiguous transmural LGE areas was significantly higher in the VT group. Using LGE-CMR, Tassi et al. (8) evaluated 59 patients with ChHD in earlier stages of myocardial involvement (preserved or minimally reduced LVEF) who had a treadmill exercise test and 24-h Holter monitoring performed in the previous 12 months. LGE was detected in 27 (45.8%) patients with a mean amount of 15.02 g. Electrical instability was detected on Holter monitoring or exercise testing in 19 patients (32%). Of these, 13 had LGE and 15 had wall motion abnormalities. It is noteworthy that electrical instability included not only sustained VT but >30 ventricular extrasystoles per hour and nonsustained VT, which could have reduced the positive and negative predictive values. Nevertheless, this study provides interesting information about the prognostic value of LGE in ChHD even in patients with preserved or minimally reduced LV function.
In this issue of the Journal, 2 studies, by Volpe et al. (9) and Senra et al. (10), present interesting new data regarding the prognostic value of LGE in ChHD. Both groups obtained follow-up data from patients with ChHD who had undergone CMR imaging to evaluate the long-term prognostic value of LGE using different endpoints. Volpe et al. (9) used a combined endpoint of cardiovascular death and sustained VT, whereas Senra et al. (10) used a combined endpoint that included all-cause mortality, heart transplantation, antitachycardia pacing or appropriate shock from an implantable cardioverter-defibrillator, and aborted sudden cardiac death. Defined secondary outcomes were the sum of cardiovascular death, sustained VT and hospitalization for cardiovascular reasons in the study by Volpe et al. (9); and all-cause mortality in the study by Senra et al. (10).
Technically, cardiac evaluation was performed by using a 1.5-T scanner in both studies, and LGE imaging acquired after 10 min in the protocol of Volpe et al. (9) and after 10 to 20 min in the protocol of Senra et al. (10). The former used the full width at half maximum method, whereas the second study used the 5 SD threshold scheme to define the extension of myocardial fibrosis.
In the study by Volpe et al. (9), 100 (71.4%) of 140 patients with ChHD showed LGE, with a median scar mass of 10.4 g representing 9.2% of the myocardium. This high prevalence of LGE was observed despite most patients being in NYHA functional class I (75%). Median follow-up was 2.75 years, during which only 14 (10%) and 31 (22%) patients reached the primary and secondary event, respectively. Presence of LGE was significantly associated with the primary and secondary event. However, in the multivariable analysis, LGE emerged as an independent predictor for mortality and sustained VT but not for the combined effect of mortality, sustained VT, and cardiovascular hospitalizations. For the latter, age and LVEF were the only independent predictors. This discrepancy was interpreted by the authors as LGE being probably related to re-entry pathways that facilitate sustained VT and sudden cardiac death, whereas LVEF is more related to the development of heart failure, which is probably true. It is noticeable that the rate of events was low and significantly inferior than in the Rassi cohort (annual mortality rate of 2% vs. 4%); thus, the results from the multivariate analyses, which included 5 independent variables, should be taken with caution due to likely model oversaturation. Although it would have been desirable to assess if the presence of LGE provided any incremental predictive value over the Rassi score, the low event rate limited further analyses. Remarkably, only 1 cardiovascular death was recorded among patients without LGE, despite low LVEF in one half of them, highlighting again the high negative predictive value of LGE.
Senra et al. (10) assessed 130 patients with ChHD. Similarly, despite most patients being in NYHA functional class I, myocardial LGE was detected in 76.1% of the population, with a mean mass of 15.2 g corresponding with 5.6% of the myocardium. Median follow-up was 5.4 years, which is almost double that reported in the study by Volpe et al. (9). As a consequence, there was a significantly higher proportion of patients who died (n = 45 [34.6%]) or reached the primary endpoint (n = 58 [44.6%]). Also contrary to Volpe et al., a significant percentage of patients had a cardiac defibrillator implanted during follow-up. LGE was significantly more prevalent and extensive in the group of patients who reached the primary endpoint. The authors additionally derived a cutoff value for LGE of 12.3 g from the receiver-operating characteristic curve that showed an area under the curve of 0.79 for the primary outcome. Remarkably again, no patient without LGE had an event during follow-up. On multivariate analysis, LGE emerged as an independent predictor of the combined endpoint both as a continuous variable and as a categorical variable. However, in this case, LVEF by CMR was not included as a covariate because, according to the authors, there was collinearity between LGE and LVEF. When the models were built including either LGE or CMR-LVEF, there were no differences in performance, which underlines the high concordance between LV systolic function and LGE for global prognosis. They did report an independent association between LGE and the primary endpoint after adjusting by the Rassi risk score, which included LV dysfunction by echocardiography as a binary variable. Contrary to Mello et al. (7), the extension of LGE was significantly larger in patients with arrhythmic events, for a similar LVEF and Rassi risk score.
The results of both studies (9,10) are relevant and in line with previous studies assessing the prognostic value of LGE in other cardiomyopathies. LGE behaves as a clear marker of poor prognosis but is highly related to other parameters of LV performance, specially LVEF. However, there is one peculiarity in patients with ChHD, namely the high incidence of arrhythmic events, and here it appears that LGE may be superior to LVEF. Larger studies with longer follow-up are needed to confirm the incremental prognostic value of LGE over LVEF to predict sustained VT and sudden cardiac death in ChHD. The most clinically relevant message from these 2 studies may be the high negative prognostic value of LGE regarding all hard cardiovascular events in ChHD.
Some other limitations should be taken into account before the routine evaluation of myocardial LGE in ChHD. First, both studies (9,10) excluded patients with pacemakers or defibrillators, who represent the most severe spectrum of the disease; second, follow-up was not scheduled but based on the retrospective review of patients’ records; and third, the lack of consensus about the best method for LGE extension quantification is relevant also because they may incur in different results (11). Finally, improvement in risk stratification in ChHD is basically useful in the decision process for device implantation, and it is not negligible that both the availability of CMR imaging and implantable cardioverter- defibrillators has obvious financial limitations in most countries where the disease is endemic.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Dr. García-Álvarez has reported that she has no relationships relevant to the contents of this paper to disclose.
- Bocchi E.A.,
- Bestetti R.B.,
- Scanavacca M.I.,
- Cunha Neto E.,
- Issa V.S.
- Weng Z.,
- Yao J.,
- Chan R.H.,
- et al.
- Rochitte C.E.,
- Oliveira P.F.,
- Andrade J.M.,
- et al.
- Volpe G.J.,
- Moreira H.T.,
- Trad H.S.,
- et al.
- Senra T.,
- Ianni B.M.,
- Costa A.C.P.,
- et al.