Abstract
The new nonlinear and dynamic assessment perspectives provided by the entropy of heart rate variability (HRV) and skin sympathetic nerve activity (SKNA) help to distinguish differences in autonomic nervous system (ANS) patterns between intradialytic hypotension (IDH) patients and non-IDH patients. These metrics have significant potential for use in clinical monitoring for hemodialysis (HD) patients.
Introduction
Hemodialysis is a life-saving procedure for patients with end-stage renal disease, but it acts as a massive stress test for the cardiovascular system. As fluid is removed, the Autonomic Nervous System (ANS) must increase heart rate and constrict blood vessels to maintain stable blood pressure. When this regulatory mechanism fails, patients experience Intradialytic Hypotension (IDH), which can lead to dizziness, organ ischemia, and increased mortality.
Traditional ANS monitoring—typically limited to linear time-domain or frequency-domain analysis of heart rate—often misses the complex, "chaotic" nature of biological signals. This study addresses this gap by combining nonlinear entropy measures (to capture complexity) with direct SKNA recording (to measure sympathetic "firing") to better predict which patients are at risk of crashing during dialysis.
Methods
The researchers enrolled 93 maintenance hemodialysis patients (66 in the IDH group and 27 in the non-IDH group). A custom-built, high-frequency wearable device was used to simultaneously record ECG and SKNA signals at a 4 kHz sampling rate during a 4-hour dialysis session.
The technical approach involved a multi-dimensional analysis:
- Nonlinear Entropy: Approximate Entropy (ApEn), Sample Entropy (SampEn), and Fuzzy Measure Entropy (FuzzyMEn) were applied to HRV data to quantify the regularity of heartbeats; lower entropy signals a less adaptable regulatory system.
- SKNA Processing: The high-frequency SKNA signal (500–1000 Hz) was rectified and averaged to calculate "average SKNA" (aSKNA), providing a second-by-second view of sympathetic activity.
- Dynamic Indices: The team calculated the difference between the start and end of dialysis (DLF) and extreme fluctuations during the session (EOF) to capture the dynamic response of the ANS.
Key Insights
- The "Broken" Correlation: In stable patients, sympathetic activity (aSKNA) was positively correlated with heart rate. In IDH patients, this correlation vanished, suggesting the heart no longer responds appropriately to sympathetic stimulation.
- The Hypertension Paradox: Patients in the IDH group often started with higher initial systolic blood pressure (average >150 mmHg). This suggests these patients may have over-activated sympathetic nerves at baseline that eventually fail or "exhaust" as dialysis progresses.
- Entropy as a Discriminator: IDH patients showed lower entropy (higher regularity) in their heart rhythms, indicating their ANS is less flexible and less capable of adjusting to the stress of fluid removal.
Results
The study identified four independent predictors of IDH: initial SBP, heart rate change (DLF), changes in the low-to-high frequency ratio (LF/HF-DLF), and extreme sympathetic nerve activity fluctuations (aSKNA-EOF).
Integrating these factors into a binary logistic regression model achieved a predictive accuracy of 84.9%. While HRV entropy was useful for distinguishing patient groups, SKNA provided a more direct and sensitive measure of the sympathetic failure that precedes or accompanies a blood pressure drop.
Conclusion
By utilizing nonlinear entropy and direct sympathetic nerve monitoring, clinicians can gain a more nuanced understanding of autonomic failure patterns in dialysis patients. Future work aims to integrate these algorithms into real-time wearable monitors, providing early warnings to medical staff before a patient's blood pressure reaches critical levels.
Summary updated