Detecting subclinical alterations in early stagesof Chagas´disease through Holter

D. S. ANDRES; I. M. IRURZUN; M. M. DEFEO; E. E. MOLA; J. MITELMAN; L. GIMENEZ; G. RANCHILIO; P. VELAZCO; J. ARMENTANO

CIRCULATION

LIPPINCOTT WILLIAMS & WILKINS

Año: 2008 vol. 117 p. 97 - 98

0009-7322

Introduction 30% of the people infected with Chagas develop symptomatic cardiac disease, including heart failure and potentially lethal arrhythmia. Following infection there exists an asymptomatic period of about 15 years, called indeterminate period (IP). Currently a method to predict which patients in IP will evolve to cardiopathic stages is not available. Objective The purpose is to test a new HRV measure (False Nearest Neighbor Fraction at dimension 10, FNNF10) for the identification of early subclinical cardiac alterations in Chagas’s disease and compare it to standard HRV indexes (RMSSD and pNN50). Methodology We studied 47 patients with Chagas’s disease and compared them to 59 healthy individuals used as control group (CG). All patients were diagnosed and classified in three groups: indeterminate period (IP - n=22), class A (A - arrhythmia or electrical abnormalities without structural heart disease, n=16) and class B (B - including dilatation and structural heart disease, n=9). 24h Holter recordings were taken by using a DMS300–7 digital recorder. We then constructed RR interval time series. We calculated standard HRV indexes (RMSSD and pNN50) and the recently developed FNNF10. The last results from applying the false nearest neighbor method (FNN), which was designed to find the embedding dimension of a time series, and from considering the value of FNN at dimension 10. It is not a statistic index, but a nonlinear measure. Results RMSSD discriminated between control and patients of either class A or B (p0.001), and showed moderated discriminative power between IP and classes A or B (p<0.02). It did not distinguish between CG and IP (p0.66) or between different stages of chagasic chronic disease (A vs. B p<0.92). pNN50 did not show discriminative power for any of the groups studied (CG vs. B p<0.11, CG vs. A p<0.41, CG vs. IP p<0.66, IP vs. B p<0.39, IP vs. A p<0.77, A vs. B p<0.66). The FNNF10 allowed to discriminate between class A or B and CG (p<0.0001) and also showed a slight difference between IP and class B (p<0.09). It did not show any differences between A and B groups (p<0.37) or IP and A (p<0.38). It showed a moderate difference between CG and IP (p<0.04). Significantly, the dispersion in the IP was high, which shows a disparity in behavior in these patients. This is in agreement with the different possible evolutions of the disease.Conclusion Between the RMSSD, pNN50 and FNNF10, the last was the only tested index able to show some difference between IP and control individuals. Among IP patients it showed high dispersion. This fact can help to identify patients whose behavior is more similar to individuals of A or B classes than to healthy ones. They could have a higher risk of developing cardiopathy. Follow up of these patients is necessary to further test this hypothesis.