Our data demonstrate that patients with epilepsy and PSPEA have a higher frequency of early developmental lesions, especially of vascular origin, and these lesions more often involve the thalamus as compared to a group of patients without PSPEA.
Cases and controls presented similar clinical and epidemiologic features. The higher frequency of cerebral palsy in cases may reflect the higher frequency of early developmental lesions in this group and may be considered more as an expression of the structural lesion of the brain than an independent factor predisposing to PSPEA.
The exact mechanism of PSPEA during non-REM sleep is unknown but the main hypothesis involves disruption of the cortico-thalamic neuronal network. The cyclical interaction between glutaminergic excitatory thalamocortical neurons in the dorsal thalamic nuclei and inhibitory GABAergic reticular thalamic neurons, located in the reticular nucleus, is the basis for the oscillating properties of the thalamus.6,8
A functional or structural lesion may lead to a switch from physiologic oscillations to pathologic epileptic discharges that may appear as generalized spike and wave complexes over the cerebral cortex.6,9,10
While the oscillating properties of this thalamic reverberating circuit are tonically inhibited by the input of the reticular activating system during wakefulness, the lack of input through the reticular activating system activity during non-REM sleep is thought to release the reverberating properties of this neuronal network.6,11,12
Several patients with thalamic lesions, in particular strokes, have developed EEG patterns consistent with ESES.13–17
In a series of 32 patients with early acquired lesions involving the thalamus ESES or PSPEA developed in 29 cases (90.6%).13
Additionally, early developmental lesions and thalamic lesions were frequently seen in patients with ESES: in a series of 67 patients with ESES, 33 patients (49.3%) had early developmental lesions,18
and in a series of 44 patients with ESES, 18 patients (41%) had early developmental lesions.19
In our series, the frequency of early developmental lesions in cases fell within the previously reported 41%–49.3% range. Furthermore, our results provide the answer to an unresolved question in previous literature: whether patients with PSPEA (cases) have a higher frequency of early developmental lesions than patients with interictal epileptiform discharges without PSPEA (controls). Our results demonstrate that cases have a higher frequency of early developmental lesions than controls, and this difference becomes even more marked when considering patients with early developmental lesions that involve the thalamus.
Because epileptiform activity occupying a minimum of 85% of the non-REM sleep tracing is by definition the most important element for the classic diagnosis of ESES,5,20
we aimed to detect whether there were imaging differences in the subgroups of cases and spike percentage during sleep above and below 85%. The higher frequency of thalamic lesions in cases with PSPEA ≥85% underlines the importance of thalamic involvement for development of the most extreme forms of PSPEA.
Early developmental lesions related to ESES are mainly of vascular origin. The first 2 case reports of early developmental lesions and ESES were described in the setting of thalamic hemorrhages.14,16
Additionally, in another series 29 out of 32 patients (90.6%) had ischemic and/or hemorrhagic prenatal or perinatal insults (stroke, infarction, periventricular leukomalacia, and/or hypoxic-ischemic encephalopathy).13
In a series of 18 patients with ESES and early developmental lesions, 14 (77.8%) had a vascular and/or infectious neonatal injury and the remaining 4 had malformations of cortical development.19
Vascular lesions were also the most frequent imaging findings (46.2%) in a series of 13 patients with surgically treated ESES and early developmental lesions.21
In another series of 33 patients with ESES and early developmental lesions, congenital stroke was the second most frequent imaging finding (21.2%) after cortical dysplasia (24.2%).18
Our data support the notion that early life strokes and periventricular leukomalacia are the main findings related to the subsequent development of PSPEA. The control group allowed us to expand these findings and demonstrate that vascular lesions are more frequent in patients with PSPEA than in a comparable population of patients with interictal epileptiform discharges during wakefulness without PSPEA.
The importance of prematurity or familial history of seizures has not been systematically studied. Based on our dataset we did not find an important contribution of prematurity or familial history of seizures to the development of PSPEA. In fact, family history of seizures was lower in cases than in a comparable population of controls. This is in accordance with previous observations that “in general, genetic factors seem to play a minor role in ESES syndrome.”5
Our data shed light on the etiology of PSPEA by demonstrating a higher frequency of early developmental lesions, especially of vascular origin, in patients with PSPEA. This finding may fuel further speculation that early disruption of the cortico-thalamic circuitry can contribute to the development of PSPEA in selected patients. Our results set the ground for future prospective studies that should follow patients with early developmental lesions and quantify the risk of developing PSPEA and ESES. The impact of our findings on immediate clinical practice is that a lower threshold for performing an MRI should be considered in pediatric patients with PSPEA as they have a higher frequency of early developmental lesions.
Our results need to be interpreted in the clinical context of the data acquisition. Data collection at a tertiary referral center may have selected a subpopulation of patients with a more severe presentation, and this may limit the overall generalizability of our results to less severely affected patients. The addition of a control group allowed us to compare patients with similar baseline characteristics except for the presence or absence of PSPEA, and therefore selection bias may not significantly affect the results as the control group has also been selected from the same tertiary referral center population.
Retrospective data collection may have been subject to information bias. We tried to overcome this bias by reviewing the original EEG tracings and MRI scans in order to standardize the data acquisition from the original studies and to reduce observer variability.
Following the basic and clinical research literature, we considered thalamic lesions as the main factor leading to PSPEA and, therefore, we focused our study on considering the presence and type or absence of early developmental lesions and the presence or absence of thalamic involvement. Other factors that can also contribute to PSPEA such as specific age at insult, location, and extension of the early developmental lesions were not specifically studied.
Selecting patients for a study on PSPEA based on their clinical presentation only without considering their EEG findings may also limit generalizability. Inclusion criteria in previous literature considered clinical and EEG findings. ESES is an EEG pattern with different clinical presentations,22
but all of them have in common an age-specific appearance of regression in, at least, one aspect of development. A proportion as high as 40%–50% of patients with ESES or PSPEA had early developmental lesions18,19
and a high proportion of patients with early developmental lesions affecting the thalamus developed ESES or PSPEA,13,14,16
but the question of whether patients with PSPEA have a higher frequency of early developmental lesions than comparable patients without PSPEA remained unanswered. We considered that the best way to obtain a control group for patients with PSPEA was to gather a population of patients with a similar clinical presentation and then determine sleep potentiation. An alternative approach of our study enrolling only patients with EEG-proven PSPEA or ESES would not have provided a control group, leaving the main question of our study unanswered.
The classic definition of ESES requires a generalized spike-wave index of at least 85% of the slow wave sleep tracing.5,20
Our study considered this definition, but also looked at patients with lower percentages of sleep potentiation. The 85% threshold is an arbitrarily set value in the initial definition of ESES5,20
and it has been largely followed.5,23–25
However, the International League Against Epilepsy (ILAE) criteria do not provide a cutoff value and only require “continuous diffuse spike-waves during slow wave sleep.”3
Other authors also consider that a lower percentage may also be consistent with ESES.26,27
We decided to include this cutoff value, but not to exclude patients with lower percentages in order to include patients along the continuum of this entity according to the ILAE definition.3
Comparing patients with and without PSPEA, a basic mechanism with different intensity of expression may shed additional light on the basic pathophysiology leading to ESES than comparing patients above and below an arbitrary threshold. In order to permit comparability with classic studies, we provided a comparison in the subgroup of cases comparing patients with a spike percentage during non-REM sleep above and below 85%.
In order to study a homogeneous population we have excluded patients with multifocal origin of epileptiform discharges, but a similar study of early developmental lesions in patients with multifocal PSPEA is under way.
Our series demonstrate that patients with PSPEA present with a higher frequency of early developmental lesions, and a higher frequency of thalamic lesions is seen in these patients than in those without PSPEA. The most frequent early developmental lesions were strokes. Early developmental lesions, especially those affecting the thalamus, and particularly in the setting of an early stroke, may play a role in the development of PSPEA.