Editor's Choice

Seizure 2022, Vol 94, Editor’s Choice: Risk of seizure recurrence in people with single seizures and early epilepsy – Model development and external validation

The unpredictability of seizures is one of the particularly disabling characteristics of epilepsy. Perhaps it is therefore not surprising that helping patients to manage uncertainty is a key aspect of the role of clinicians working in this field. Uncertainties are a central feature in many clinical encounters involving patients with epilepsy – in discussions about the causes of epilepsy, the meaning of investigation findings, the likely benefits and drawbacks of treatment, the risks asssociated with different activities or longterm outcomes. Indeed, health professionals who struggle to embrace uncertainty should probably stay clear of a career in epileptology.

There are, of course, many things about epilepsy we know. For instance, we can be quite certain of the range and contribution of different aetiological factors, reliability of investigation findings, chances of achieving seizure control and longterm outcomes at population and or even at specific patient cohort level. Our challenge is to bridge the gap between this knowledge and the particular patient – for instance, a patient who needs to decide whether to start antiseizure medication (ASM). Experts tasked with helping individuals in this clinical scenario have evidence from many studies at their disposal (1), but the integration of this evidence and its application to the personal circumstances of the patient in the room will draw more on the art than the science of medicine.

The recent ILAE decision that clinicians should diagnose epilepsy after a single seizure in the presence of a 60% risk of seizure recurrence has not really helped with decisions about starting antiseizure treatment – it has only brought the need to make a decision into starker relief. It has also introduced a numeric threshold (without taking account of the size of the risk per year) although the precise lifetime seizure recurrence risk is impossible to calculate accurately for an individual patient.

One important step to reducing the gap between diverse bits of evidence and individual treatment decisions involves statistical modeling. This approach can deliver numeric risk values and take account of a range of potentially relevant clinical characteristics of the patient in the room. My Editor’s Choice from the current volume of Seizure is an article by Laura Jayne Bonnett which validates one such model (2). This paper follows the ‘gold standard’ modeling approach of externally validating a model developed on the basis of data from one dataset (the large MESS study) (3) by testing the model on other, independently collected datasets (in this case from the equally impressive NGPSE, Western Australia and FIRST studies) (4-6). My Editor’s Choice demonstrates that, in patients with a single or small number of seizures first seeking neurological advice, the model is capable of correctly predicting the likelihood of seizure occurrence by one or three years. The baseline mean estimate of seizure recurrence were 35.1% and 46.2% at these time points, but the model will demonstrate its real clinical utility when the patient’s particular personal risk factor profile is used to generate an individualised risk estimation.

Admittedly, there is more work to do (for instance in terms of finding out what sense patients make of a given percentage of seizure recurrence risk). However, providing patients with their individual risk of having another seizure within one or three years should certainly help to close the gap between knowledge from research studies and personal decisions about starting ASMs.


(1) Rizvi S, Ladino LD, Hernandez-Ronquillo L, Téllez-Zenteno JF. Epidemiology of early stages of epilepsy: Risk of seizure recurrence after a first seizure. Seizure 2017;49:46-53.

(1) Bonnett LJ, Kim L, Johnson A, Sander L, Lawn N, Beghi E, Leone M, Marson A. Risk of seizure recurrence in people with single seizures and early epilepsy – model development and external validation. Seizure 2022; 94: 26-32

(2) Marson, A., et al., Immediate versus deferred antiepileptic drug treatment for early epilepsy and single seizures: a randomised controlled trial. Lancet 2005; 365(9476): 2007-2013.

(3) Hart, Y.M., et al., National General-Practice Study of Epilepsy - Recurrence After a First Seizure. Lancet, 1990. 336(8726): p. 1271-1274.

(4) Lawn, N., et al., Is the first seizure epilepsy—and when? Epilepsia 2015;56: 1425-1431.

(5) Musicco, M., et al., Randomized Clinical-Trial on the Efficacy of Antiepileptic Drugs in Reducing the Risk of Relapse after a 1st Unprovoked Tonic-Clonic Seizure. Neurology 1993;43:478-483.