Author + information
- Published online September 18, 2017.
- aDepartment of Neurology, Leiden University Medical Centre, Leiden, the Netherlands
- bStichting Epilepsie Instellingen Nederland (SEIN), Heemstede, the Netherlands
- ↵∗Address for correspondence:
Dr. J. Gert van Dijk, Department of Neurology, Leiden University Medical Centre, PO Box 9600, 2300 RC Leiden, the Netherlands.
In this issue of JACC: Clinical Electrophysiology, Sahota et al. (1) present a very creative way of looking at the recurrence of vasovagal syncope (VVS), in patients with very frequent and refractory attacks. They studied the duration of intervals between days on which VVS occurred, and used a Poisson distribution to deduce that VVS occurred randomly in time.
This may seem surprising. VVS is just 1 of many paroxysmal diseases, and compared with some others, notably epilepsy, VVS differs in that it is very often triggered, whereas epileptic seizures mostly seem to occur out of nothing. In fact, the specific triggers evoking VVS are often the reason to diagnose a specific event as vasovagal in nature: when someone briefly loses consciousness during venipuncture, the list of disorders to choose from effectively consists of VVS only. This reliance of VVS on triggers seems at odds with a random occurrence, but in this paper randomness does not refer to a lack of triggers, but rather to a specific temporal pattern.
That pattern is the Poisson distribution. It is very distinct from the distribution physicians are probably more familiar with, the “normal” one, with its nice symmetrical distribution around a mean that makes it easy to calculate threshold values. If some event occurs randomly over time, the intervals between successive events fit the Poisson distribution: its distribution is skewed, with many more short intervals than long intervals. Its shape is very well illustrated by Figure 2 in the paper (1). This skewed nature can be grasped by a thought experiment: take a yardstick and carve a notch in it somewhere at random, dividing the stick into 2 lengths. Adding a new cut at random will divide 1 of the previous lengths into 2 shorter ones, and so on. It is unlikely that a really long length will remain intact for long, explaining the rarity of long intervals. In Sabota et al.'s (1) study, the VVS distribution resembled a Poisson distribution very closely, so an underlying random process seemed very likely.
A clear trigger mechanism may still seem at odds with a random time distribution in VVS, but there may not really be any contradiction. If, theoretically, VVS would only occur after a specific trigger, but always following that trigger, than the distribution of VVS would simply be the distribution of the trigger, whatever its nature. But such an unrelenting and obligatory trigger is very unlikely in VVS. Instead, the patients in this study probably had multiple triggers as well as many putative trigger instances. The latter factor may be important: although standing is a strong trigger of VVS, the periods of standing that do not evoke VVS are much more common than are the periods in which it does, even in people with frequent VVS. This makes it likely that some underlying process modulates the likelihood of VVS; the triggers provide a glimpse at the activity of that process. This underlying process may be the one that critically determines the shape of the distribution, much more so than the occurrence of the triggers.
Of course, there may be more than 1 such underlying process at work. It is also conceivable that suffering a VVS spell temporarily decreases the chance of having another one, or in reverse the spells may temporarily increase that likelihood. In epilepsy, the existence of “kindling” has long been debated (2), which reflects such a positive feedback mechanism. Sahota et al. (1) present very suggestive material that there are multiple underlying processes at work in VVS: although some people may have just 1 such process going on, others may have multiple ones, acting at the same time or after one another, resulting in changes in the rate of VVS, upward as well as downward.
The authors wisely do not speculate on the nature of these processes. They may include subtle changes in fluid and salt intake. Based on patients' stories, we have often suspected that anxiety and stress are important long-term modulators of the VVS rate; in some cases, the anxiety caused by VVS itself may increase the tendency to respond with a VVS response on a short time scale. But for now, all this remains speculation.
Sahota et al. (1) introduced the tools of temporal distribution of paroxysmal events, previously used in a few other paroxysmal disorders, to VVS research. Given the large number of people with VVS and the high frequency of VVS in many patients, such analyses are likely to work well in VVS. The resulting lessons may well be put to good use in epidemiological and treatment studies. Ultimately, they may help answer another question, buried several layers underneath the present study: why do humans have this odd reflex, fired by innocuous physiological events, that turns off the circulation and hence the brain (3)?
↵∗ Editorials published in JACC: Clinical Electrophysiology reflect the views of the authors and do not necessarily represent the views of JACC: Clinical Electrophysiology or the American College of Cardiology.
Dr. van Dijk has reported that he has no relationships relevant to the contents of this paper to disclose. Dr. Thijs has received research support from ZonMW, the Dutch Epilepsy Foundation, the NUTS Ohra Fund, Medtronic, and the AC Thomson Foundation; and fees for lectures from Medtronic, UCB, and GlaxoSmithKline.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- 2017 American College of Cardiology Foundation