Abstracts

Rationale: From its roots in studying cardiac autonomic instability, heart rate variability (HRV) has received attention as a potential biomarker for vagal nerve stimulation (VNS) efficacy. The literature on VNS effects on HRV has significant heterogeneity in study methods and evaluated outcomes, making a rigorous meta-analysis somewhat constraining. A scoping review of the literature is of benefit to provide a comprehensive look at the available data as well as to better understand the experimental conditions needed for future, better, inquiry.

Methods: A scoping literature review was conducted using PubMed, Web of Science, EMBASE, and Cochrane Library using PRISMA guidelines. Studies were included if they were primary articles, involved an implantable VNS device, had HRV as an outcome, and used human volunteers. Non-English articles were excluded, and backward citation chaining was performed as needed. From the final papers two authors abstracted demographics, VNS and ECG methodology, statistics, and study outcomes.

Results: A total of 29 studies were included. VNS settings were mostly managed according to clinical practice, rather than systematically to better illuminate parametric effects. ECG sampling rate, RR-correction and the statistical analysis of outcomes varied widely and were sometimes unreported. All studies included one or more of the following comparisons.

Pre- vs. post-implant: Increases in HF power, pNN50, and RMSDD were noted in the first-year, but these differences disappeared after 12 months. SDANN was increased during parasympathetic but decreased during sympathetic challenges post-implant.
Control vs. pre-implant: Other than a significant increase in HR during sleep and decreased non-linear variables, results among studies varied.
Control vs. post-implant: A decrease in multiple HRV measures was seen 6-12 months post-implant.
VNS-on vs. VNS-off: No significance was seen across HRV measurements, although HF and LF power had conflicting results.
VNS responders vs. non-responders: Conflicting HRV results were seen in pre-implant patients. After implant, frequency domain and RMSDD results were increased only in responders with seizure reduction >50%. By 1 year follow up, there were no significant differences in any HRV measure.

Conclusions: There are clear alterations in HRV when VNS is implanted; however, within a year no significant differences remained. We suspect that HRV migrates back to pre-implant homeostatic points. There is preliminary evidence that frequency variations and RMSDD may be a predictive marker for post-implant VNS efficacy. However, if we hope to better understand the relationship of vagal modulation on this primary oscillatory signal, more rigorous and standardized methodology must be used in future studies. The subtlety of sympathetic/parasympathetic signaling likely requires a more structured approach to experimental and analytic techniques to better elicit signal. Non-linear evaluations as well as the covariate analysis of somatic and cerebral oscillatory signals in humans might provide a more robust biomarker for VNS efficacy.

Funding: None