Paper accepted in Bioengineering
Our latest study on heart rate variability for the prediction of delayed cerebral ischemia has been published in Bioengineering. Delayed cerebral ischemia (DCI) affects up to a third of patients after aneurysmal subarachnoid hemorrhage (aSAH), and established risk scores are static — they cannot capture the dynamic physiological changes that precede the event. In this study, we asked whether heart rate variability (HRV), a continuous non-invasive marker of autonomic function, adds information beyond the clinical and laboratory predictors used in our previous longitudinal DCI model.
We developed an algorithm for the real-time computation of time- and frequency-domain HRV features from high-resolution ECG and evaluated it in 101 patients with aSAH, 35 of whom developed DCI. Among individual HRV features, normalized low-frequency power (LFNorm) was the strongest univariate predictor (ROC AUC 0.64). Combining HRV with static clinical variables and laboratory/blood gas data yielded the best overall performance (ROC AUC 0.68), outperforming each modality alone, with HRV accounting for the largest share of feature importance. Time-resolved analyses showed that the contribution of each modality shifts over the clinical course, with HRV providing its greatest incremental value in the later phase preceding DCI.
Read the full article here: https://doi.org/10.3390/bioengineering13080857