The aim of this study was to assess whether heart rate variability (HRV) could predict which hemisphere of the brain was affected during an acute ischemic stroke (AIS). To achieve this goal, we compared HRV between patients with a right (RH) and left hemispheric (LH) stroke. The studied group consisted of 64 patients with AIS (25 with RH and 39 with LH stroke, with a mean age of 64 ± 12 and 66 ± 13, p = 0.3, respectively) using 24 h Holter ECG records at NN intervals performed at a mean of 4.3 ± 2 days following their AIS. Standard linear methods were used to analyze HRV in the time and frequency domains, as well as nonlinear methods, including sample entropy, detrended fluctuation analysis, and asymmetry measures. Patients with an LH stroke had significantly greater values for sample entropy compared to subjects with an RH stroke (1.31 ± 0.53 vs. 0.92 ± 0.46, p = 0.003, Bonferroni-corrected p = 0.033, effect size = 0.8). The LH stroke group also had higher RMSSD (113 ± 81 vs. 76 ± 61, p = 0.06), pNN50 (33.35 ± 28.54 vs. 18.52 ± 23.75, p = 0.02), and HFnu (48.42 ± 16.41 vs. 42.66 ± 17.88, p = 0.11) values, when compared to the RH group, which was possibly related to higher activity in the parasympathetic system in the LH group. Conversely, subjects with RH stroke had higher LFnu (57.34 ± 17.88 vs. 51.58 ± 16.41, p-value = 0.11) and LF/HF ratios (2.24 ± 2.87 vs. 1.68 ± 2.50, p-value = 0.11), which were likely related to higher activity in the sympathetic nervous system, when compared to the LH stroke group. Our pilot study demonstrated that patients with RH stroke had lower HRV complexity than those with LH stroke, indicating that HRV could be useful in the discrimination of hemispheric involvement in AIS.
Keywords: brain hemisphere; complexity; heart rate variability; ischemic stroke; nonlinear analysis; sample entropy.