Influence of type 1 diabetes on the symbolic analysis and complexity of heart rate variability in young adults

Abstract

Background: Type 1 diabetes mellitus can cause autonomic changes, which can be assessed by heart rate variability. Among the heart rate variability assessment methods, the symbolic analysis and Shannon entropy, based on the Chaotic dynamics, have gained prominence. Objective: To compare heart rate variability indexes, obtained through symbolic analysis and Shannon entropy, in young adults with type 1 diabetes mellitus and healthy young individuals, associated with the analysis of linear indexes; and to verify if there are associations between the indexes obtained by the symbolic analysis and by Shannon entropy and linear indexes in diabetic individuals. Methods: Heart rate variability data from 39 young adults with type 1 diabetes mellitus and 43 healthy young individuals were analyzed, using a cardio-frequency meter. Linear indexes (standard deviation of all normal RR intervals recorded in a time interval expressed in milliseconds; square root of the mean of the squared differences between adjacent normal RR intervals in a time interval expressed in milliseconds; low and high frequency components in millisecond squared; and normalized units and ratio between low and high frequency components) and nonlinear ones (Shannon entropy and symbolic analysis – standard without variation; with one or two variations; and with two different variations) of the heart rate variability were calculated. The statistical significance was set at 5%, and the confidence interval was 95%. Results: Significantly lower values were observed in the DM1 group compared to healthy young adults for the standard deviation indexes of all normal RR intervals recorded in a time interval [37.30 (29.90) vs. 64.50 (36.20); p = 0.0001], square root of the mean of the squared differences between adjacent normal RR intervals in a time interval [32.73 (17.43) vs. 55.59 (21.60); p = 0.0001], low frequency component [402.00 (531.00) vs. 1,203.00 (1,148.00); p = 0.0001], high frequency component [386.00 (583.00) vs. 963.00 (866.00); p = 0.0001] and the pattern with two different variations [15,33 (9,22) vs. 20.24 (12.73); p = 0.0114], with the effect of this difference being considered large (standard deviation of all normal RR intervals recorded in a time interval, square root of the mean of the squared differences between adjacent normal RR intervals in a time interval and low frequency component), medium (high frequency component) and small (standard with two different variations). The agreement of the associations between the linear and non-linear indexes was considered elevated for the high frequency component index - normalized units (r = –0.776), with the standard index without variation, and moderate for the indexes square root of the mean of the squared differences between adjacent normal RR intervals in a time interval (r = 0.550), standard deviation of all normal RR intervals recorded in a time interval (r = 0.522), high frequency component - normalized units (r = 0.638) with the index standard with two similar variations, as well as for the indexes square root of the mean of the squared differences between adjacent normal RR intervals in a time interval (r = 0.627) and high frequency component - normalized units (r = 0.601) with the index standard with two different variations. Conclusion: Type 1 diabetes mellitus influenced linear indexes and symbolic analysis, but not yet in the complexity of heart rate variability. Additionally, heart rate variability indexes correlated with the symbolic dynamics.