Importância da variabilidade da frequência cardíaca e dos gráficos de recorrência como marcadores de prognósticos em pacientes de Unidade de Terapia Intensiva

Abstract

The Intensive Care Units are designed to assist high-risk patients, performing uninterrupted care by a multidisciplinary team with appropriate equipment. These units demand increasing costs, requiring specific data on the severity and prognosis of patients, enabling interventions by the health team, both for assistance and resource planning. Thus, defining a prognostic marker in this setting is of utmost importance OBJECTIVES: To evaluate patients admitted into the ICU by different diagnoses, considering the changes in their clinical status and their respective autonomic changes, comparatively in the different domains of Cardiac Frequency Variability. To associate HRV variables with higher probability of death and/or worse prognosis in a short term. METHOD: For HRV analysis, the collection of time series of heartbeats was used with the aid of the Polar Advanced RS800CX frequency meter, for about 30 minutes. A total of 92 patients were analyzed. Data were screened and subsequently analyzed with the Kubios HRV Analysis software. Time, frequency and non-linear variables were selected. A survey of clinical data was carried out from the period of hospitalization to discharge. RESULTS: In the death group, 50% were smokers and 68.8% were hypertensive, while comorbidities as a cause of hospitalization; 34.8% were due to respiratory diseases. In the death group, males were predominant (68.8%), and more deaths in patients on mechanical ventilation (78.1%). Significant difference in heart rate variability was found between death versus time-domain survivors [SNS Index (6.7 ± 4.2; 4.7 ± 3.9; P = 0.0257), Stress Index (40.6 ± 20.7; 29.9 ± 19.7; P = 0, 0160), SDNN (8.4 ± 10.4; 15.4 ± 13.1; P = 0.011), RRTRI Index (2.8 ± 3.1; 4.4 ± 3.4; P = 0.0345 ) and TINN (54.8 ± 55.9; 92.2 ± 73.9; P = 0.0126)], in the frequency domain [VLFm² (13.2 ± 26.4; 35.5 ± 63.1 ; P = 0.0092), LFm² (63.4 ± 152.7; 193.9 ± 358.1; P = 0.0012), HFm² (55.1 ± 194.1; 135.3 ± 294.4 ; P = 0.0004) and Total Power m² (132.3 ± 357.3; 365.2 ± 685.6; P = 0.0006)] and in the non-linear domain [SD1 (6.5 ± 9 , 7; 10.7 ± 10.4; P = 0.001), SD2 (9.7 ± 11.3; 18.5 ± 15.8; P = 0.0004)], respectively. The analysis by means of Recurrence Graphs did not show differences between the groups by the quantitative analysis, while the qualitative analysis suggested a greater presence of geometric shapes, increased plot and dark squares; all pointing out a greater linearity in the group of deaths. CONCLUSION: We can conclude that HRV with the aid of Recurrence Graphs provides differentiation on severity according to health states, and can work as a prognostic marker both in the improvement of clinical status or probable death.