Comparison between symbolic and spectral analyses of short-term heart rate variability in a subsample of the ELSA-Brasil study.

Linear and nonlinear analyses of heart rate variability (HRV) have been largely used to evaluate the autonomic balance directed to the cardiovascular system. However, comparative studies evaluating the agreement between methods are scarce. Therefore, our aim was to examine the relationship between spectral (SPA; linear) and symbolic analyses (SYA; nonlinear) indexes. A subsample of 683 participants of the Brazilian Longitudinal Study of Adult Health was investigated. Linear and nonlinear analyses were obtained from 10 min ECG recording at rest. Reliability and agreement between methods were evaluated by kappa-statistic and proportion of agreement. According to SYA, the most frequent pattern was P1V (sympathovagal balance, without sympathetic or vagal predominance) comprising 62.7% of the sample, followed by P2V (vagal predominance) with 33.2%, and finally P0V pattern (sympathetic predominance) with 4.1%. Overall proportion of agreement between SYA and SPA was 39.68% (95% CI 0.360–0.433), with expected agreement by chance of 30.8%. Kappa value was 0.128 indicating a slight agreement between methods. Proportion of agreement was 7.93% (95% CI 0.032–0.126) for predominant sympathetic modulation, 10.39% (95% CI 0.075–0.132) for sympathovagal modulation, and 40.29% (95% CI 0.361–0.444) for parasympathetic modulation. Our data provide evidence for important differences between SPA and SYA on HRV analysis. More studies are needed to clarify the causes of disagreement between two methods designed to evaluate the autonomic modulation of heart beats.