Introduction:
Currently, traditional markers of heart rate variability (HRV) are used as measures of sympathovagal balance, reflecting autonomic dysfunction, in heart failure. However, reference values of HRV markers and their clinical relevance is as of yet undetermined.
Methods:
For this study data from the MyoVasc cohort (NCT04064450; N=3,289), a prospective cohort study on individuals with heart failure (HF), were investigated. Study participants were classified according to the universal definition of heart failure (at risk of HF: stage A, pre-HF: stage B and heart failure: stage C/D), and received a 24-hour Holter ECG device after a highly standardized examination, including 2D echocardiography. HRV markers from the time domain, frequency domain, and non-linear indices were evaluated utilizing a systematic literature screen ranking HRV markers by search hits, and a random survival forest model predicting cardiac death. The top 10 ranked HRV markers from the literature screen and the data driven method were further investigated. Reference values were determined in a cardiovascular healthy subsample. HRV markers were investigated for their clinical relevance in multivariable Cox regression analyses considering potential confounders for their relationship with all-cause death in participants with heart failure.
Results:
Holter ECG recordings were available for analysis in 855 study participants with HF (mean age 65.6±10.1 years; female sex: 33.9%) and 133 healthy individuals (mean age 57.8±10.3 years; female sex: 46.6%). The literature screen revealed that the top 10 most frequently reported HRV markers were from the time and frequency domains, while non-linear HRV measures were amongst the top ranked markers from the data-driven approach. Multivariable Cox regression models demonstrated prognostic value of the HRV markers for all-cause death in the HF analysis sample. From the time and frequency domain the mean heart rate (hazard ratio (HR) per standard deviation [HRSD]: 1.21 [95% CI: 1.01-1.45], p=0.04), LF/HF (HRSD: 0.71 [95% CI: 0.58-0.86], p=0.0005), and total power (HRSD: 0.84 [95% CI: 0.71-0.98], p=0.03) were relevantly prognostic for all-cause death in the age and sex-adjusted models. However, none of the time and frequency domain parameters were prognostic after additional adjustments for traditional cardiovascular risk factors (CVRFs), comorbidities and medication intake. Acceleration capacity [AC] (HRSD: 1.61 [95% CI: 1.32-1.96], p<0.0001), deceleration capacity [DC] (HRSD: 0.66 [95% CI: 0.56-0.78], p<0.0001), and time lag (HRSD: 1.27 [95% CI: 1.10-1.47], p=0.001), were prognostic non-linear indices for all-cause mortality in the age and sex adjusted models as well as in the models additionally adjusted for CVRFs, comorbidities, and medication intake (AC, HRSD: 1.53 [95% CI: 1.21-1.93], p=0.0004; DC, HRSD: 0.70 [95% CI: 0.55-0.88], p=0.002; time lag, HRSD: 1.22 [95% CI: 1.03-1.44], p=0.018).
Conclusion:
HRV values outside of the reference range, and non-linear HRV markers reflecting autonomic dysfunction are independent predictors of all-cause death in individuals with heart failure. This suggests that HRV could potentially be used to improve risk stratification, and develop new strategies for intervention.