Despite the facts that wearable technology capable of monitoring a range of physiological function has improved greatly over the last decade and that the self-monitoring of physiological functions using this sort of technology is now accepted as an entirely normal thing to do, there are still many issues with the routine use of self-monitoring equipment to warn people with epilepsy about impending seizures, or at least to count seizures accurately once they have happened. While significant progress in this area has been charted in a previous Editor’s Choice article in this journal, seizure detection devices are still associated with unacceptable error rates, both in terms of under- and over-detecting seizures. What is more, different devices have different strength and weaknesses, and a device that works reliable in one patient may not work well in another (1).
My Editor’s Choice from the current edition of Seizure, a systematic review and meta-analysis of pre-ictal heart rate changes by Elisa Bruno et al. focuses on a physiological parameter which currently available wearable devices are capable of measuring well and easily: the heart rate (2). This review, based on results from a total of 1,110 people living with epilepsy and 2,957 seizures finds that pre-ictal heart rate increases (HRI) were found in 623/2,957 (21%) seizures. The median onset time of pre-ictal HRI was 10.7 seconds prior to seizure onset (IQR 5-60). The studies which could be included in the meta-analysis yielded a pooled incidence of pre-ictal HRI of 36/100 seizures (95% CI 22-50). More fine-grained analyses suggested a higher prevalence of pre-ictal HRI in temporal lobe than other focal or generalized epilepsies as well as a higher prevalence in adults than children. Not unexpectedly, ictal HRI were seen more commonly than pre-ictal HRI. The cumulative incidence of ictal HRI was 1,556/2,957 seizures (52.6%), with a mean increase in heart rate of 56.8% compared to baseline.
These findings suggest that seizure warning or detection systems including sensors capable of monitoring patients’ heart rate should be useful for a fair proportion of individuals with epilepsy, in particular for adults with temporal lobe epilepsy. However, the data also demonstrate that even perfectly sensing heart-rate based systems could only ever provide a minority of patients with seizure warnings, and that warnings could usually only be given a few seconds prior to seizure onset. Even seizure detection (rather than warning) devices based on HRI alone could only hope to pick up about one half of all seizures.
It is possible that more advanced forms of analysis of cardiac activity (such as heart rate variability analyses) (3) could do better than systems based on heart rate analysis alone, but this review and meta-analysis indicates that it is more likely that effective seizure warning or detections systems must also involve the observations or other parameters and that they will need to be individualised to particular patients’ seizure disorders.
(1) Jory C, Shankar R, Coker D, McLean B, Hanna J, Newmann C. Safe and sound? A systematic literature review of seizure detection methods for personal use. Seizure 2016;36:4-15.
(2) Bruno E, Biondi A, Richardson MP, on behalf of the RADAR-CNS Consortium. Pre-ictal heart rate changes: a systematic review and meta-analysis. Seizure 2018, please add bibliographic details.
(3) Pavei J, Heinzen RG, Novakova B, Walz R, Serra AJ, Reuber M, Ponnusamy A , Marques JLB. Early seizure detection based on cardiac autonomic regulation dynamics. Frontiers of Physiology 2017;8:765, 1-12.