JLE

Epileptic Disorders

MENU

Gelastic seizures: a retrospective study in five tertiary hospital centres Volume 22, numéro 2, April 2020

Illustrations


  • Figure 1

Tableaux

The first description of ictal epileptic laughter dates from 1873, by Trousseau (Trousseau et al., 1873). Henceforth, seizures with this manifestation were named “gelastic seizures” (GS), from the Greek word “gelos”, that means laughter. GS are characterized by recurrent bouts of paroxysmal stereotyped laughter or giggling, generally without mirth or appropriate affective tone (Gascon et al., 1971; Chen at al., 1973). It is a rare entity with an estimated prevalence between 0.4 and 0.8% (Kovac et al., 2015). These seizures can consist exclusively of laughing or be associated with general autonomic arousal, motor automatisms or disturbed consciousness (Cerullo et al., 1998).

GS are more likely to be diagnosed during childhood, classically associated with hypothalamic hamartoma (HH) (Striano et al., 2009). In these cases, they are often refractory to pharmacological treatment, and precocious puberty and cognitive impairment are frequently associated (Nguyen et al., 2003).

However, GS may also arise from other brain regions, including temporal, frontal or parietal lobes. In these cases, areas involved in the physiological network of smiling and laughter, connecting the hypothalamus with other regions, particularly the temporal and frontal lobes, should be considered (Wild et al., 2003).

Most studies of GS are dedicated to the paediatric population. Patients with GS without HH have been mostly described in very few case reports, or case series, with small sample sizes (Tran et al., 2014; Kovac et al., 2015; Gutierrez et al., 2016). These case series used video-EEG (VEEG) databases mostly from patients with refractory epilepsy based on pre-surgical work-up, hence patients were subjected to selection bias, as only severe epilepsy patients were included.

This study aimed therefore to characterize, clinically and neurophysiologically, a series of patients with GS collected from epilepsy clinics of five tertiary hospitals within a single country.

Material and methods

This retrospective study included patients with GS from epilepsy clinics from five Portuguese tertiary hospital centres: Hospital de Santa Maria (Centro Hospitalar e Universitário Lisboa Norte; Hospital de São José - Centro Hospitalar Universitário de Lisboa Central; Hospital Pediátrico de Coimbra; Centro Hospitalar Universitário de Coimbra; Centro Hospitalar do Hospital de São João).

All patients with seizures defined by the ILAEOperational Classification of Seizure Types, 2016 as GS were included, based on clinical and VEEG databases at the participating centres. Relatives’/caregivers’ descriptions and home videos were accepted for ictal semiology description. Exclusion criteria were cases with absence of interictal EEG or cranial MRI, or limited access to detailed clinical data.

From the patient clinical files, the following variables were extracted: sex, age at first GS (infancy [≤10 years old], adolescence [between 10 and 18 years old] and adulthood [≥18 years old]), GS as first manifestation of epilepsy, presence of other seizures and other conditions, aetiology and topography (based on cranial MRI), interictal and ictal (for patients with VEEG) EEG topography, therapeutics, epilepsy surgery, and follow-up after surgery.

All data were evaluated using descriptive statistics (SPSS 21).

Results (table 1)

 

Thirty-five patients with GS were obtained from the participating centres, 20 (57.1%) males and 15 females (42.9%). The first GS occurred between two months and 47 years of age, (median of six years). Twenty-two patients (62.9%) had a first seizure in infancy, five (14.3%) in adolescence and eight (22.8%) at adult age. In 15 patients (48.4%), GS were the first manifestation. The majority of patients presented also with other seizures besides GS. This information was available in 31 patients, of whom 27 (87.1%) had other seizures, such as generalized tonic-clonic seizures and focal seizures with motor or sensitive onset. Delayed psychomotor development was present in 58.1% of these 31 patients. In the HH group, all patients had concomitant psychomotor delay or cognitive deficits and one patient also had precocious puberty.

Diagnosis was made based on VEEG in 16 (45.7%) patients. For the remaining patients, ascertainment was based on descriptions by caregivers/relatives and direct clinical observation.

Twenty-six patients had abnormal MRI. Structural epilepsy was diagnosed in 24 patients: eight patients with HH and 16 with various aetiologies. One patient presented with genetic epilepsy (Niemann Pick C), and another had immune epilepsy related to Rasmussen disease. Lesions on MRI were either on the left (n=9) or on the right side (n=6). Nine patients had normal cranial MRI.

Regarding the eight patients with HH, seven (87.5%) had their first GS in infancy and one (12.5%) in adolescence. For the remaining aetiologies, 16 (59.3%) patients started having GS in infancy, three (11.1%) in adolescence and eight (29.6%) at adult age.

Interictal epileptiform activity was found in 31 patients (81.6%), including most patients with HH (7/8). Interictal epileptiform activity usually included either the frontal or the temporal lobes (n=26; 83.9%). However, five (16.1%) involved the parietal or the occipital lobe.

VEEG data was available for analysis in 11 patients. Scalp EEG ictal onset was temporal in five patients, frontal in four, parietal in one, and occipital in one. Seven of these patients had non-HH lesions on cranial MRI. Among these, six had a VEEG concordant with the MRI lesion: frontal (one patient), temporal (three patients), parietal (one patient) and occipital (one patient) (figure 1).

In 30 patients, it was possible to evaluate the treatments prescribed. From those patients, 90% (n=27) had refractory epilepsy. All the treatments are described in table 1. Epilepsy surgery was performed in 36.7% (n=11) of patients. Only three patients with HH were submitted to surgery. In the HH group, only one of the three patients submitted to surgery remained seizure-free after 14 months of follow-up. In the non-HH group, on the contrary, most patients remained in Engel Class Ia, except for one patient with a genetic disorder, for whom surgery was ineffective.

Discussion

In this study, we reported 35 adult and paediatric patients with GS and identified different aetiologies and affected topographies, including 27 patients without HH. Cases were obtained not only based on VEEG but also from other sources for seizure characterization, and patients with both laughter and giggling were included, providing a more comprehensive picture of GS in clinical practice.

In most patients, GS were accompanied by other seizures, including mostly generalized tonic-clonic seizures and focal seizures with impaired consciousness and motor or sensitive onset. These data are similar to those of other studies, which also concluded that patients with GS usually have other types of seizures (Tassinari et al., 1991; Striano et al., 1999, 2012). Most seizures are automotor seizures, that commonly originate in temporal or frontal regions, and involve the different areas involved in GS network and propagation of seizure activity. GS was the presenting symptom of epilepsy in only 15 of the 31 patients based on the information available. In both HH and non-HH, approximately half the patients had other types of seizures before GS were noted or reported.

The majority of previously published GS series have addressed HH related epilepsies (Nguyen et al., 2003). Indeed, GS are commonly deemed a hallmark of HH, especially in early-age-onset cases (Kerrigan et al., 2005). In this type of pathology, patients can present with other conditions, like delayed psychomotor development, severe cognitive deficit or precocious puberty (Tassinari et al., 1991). This is congruent with our data regarding other associated conditions. In our study, we found eight patients (22.9%) with GS associated with HH, and in all of them, seizures started during infancy or adolescence, reinforcing this classic association.

In patients without HH, the age at the first GS varied significantly, and a high proportion started in adolescence and adulthood. This is in accordance with previously published series of non-HH GS in which seizure onset ranged from between months to 93 years of age (Tassinari et al., 1991; Kovac et al., 2015; Gutierrez et al., 2016).

There are many different non-HH GS associated aetiolies. These include inflammatory, infectious, and dysplastic aetiologies, as well as mesial sclerosis and tumours of all cerebral lobes, the pituitary gland, mammillary bodies and the third ventricle (Tassinari et al., 1991; Munari et al., 1995; Kuzniecky et al., 1997). Although mesial temporal lobe sclerosis was not suggested to be an aetiology of GS based on earlier studies, (Tassinari et al., 1991), later series have shown a predominance of mesial sclerosis, ranging from 12.5 to 21% (Kovac et al., 2015; Gutierrez et al., 2016). In our series, the percentage was smaller (7.4%), suggesting that this preponderance in previous studies is most likely related to the reference bias associated with VEEG.

As previously mentioned, localization of the epileptogenic focus in GS is presumably related to the cerebral network of laughter that spreads from the hypothalamus to several cerebral areas, such as the inferior temporal, the mesial frontal and the parietal lobes (Umeoka et al., 2008; Talvik et al., 2012). Cortical stimulation studies showed widespread areas involved in that network (Caruana et al., 2015). In our series, data from both cranial MRI and interictal and ictal EEG reinforces this claim. Indeed, interictal epileptiform activity and ictal EEG onset was located at fronto-temporal leads, respectively, in 83.9% and 71.4% of the patients. This is also in accordance with previous reports of GS (Tassinari et al., 1991; Striano et al., 1999; Tran et al., 2014; Kovac et al., 2015; Gutierrez et al., 2016). In particular, Tassinari and co-workers (Tassinari et al., 1991) described the largest series of GS published to date. These authors reviewed 60 patients with GS not associated with HH and 60 patients with HH, mostly from previously published cases. In accordance with our data, in non-HH patients, seizure foci were located in sites that included mostly temporal (61%), frontal (10%), and frontal and temporal (22%) areas. Localization of seizure onset was not possible in 7% of the patients. Cortical stimulation studies also show that the temporal lobe generates smiling manifestations (Arroyo et al., 1997; Wild et al., 2003). The frontal lobe has also been shown to be involved in the motor aspect of laughter, with a non-emotionally driven pathway involving the anterior cingulate, premotor frontal, and opercular areas projecting through the motor cortex directly to the ventral brainstem (Wild et al., 2003; Tran et al., 2014). Tassinari described semiological differences between GS according to its origin and aetiology (Tassinari et al., 1991). GS with temporal lobe origin could resemble normal laughter and usually occurred with a sense of mirth, pleasure or happiness. On the contrary, frontal lobe origin (six cases) led to “unnatural” and briefer periods of laughter without emotional content that were usually less frequent with long-lasting laughter. In HH, GS resembled natural laughter but lacked emotional content and were usually brief. These semiological differences are in accordance with this proposed role of the frontal lobe in the motor component of laughter, whereas limbic structures are mostly responsible for the emotional part of this behaviour. Unfortunately, in our study, given its retrospective nature, we were unable to analyse the motor and emotional characteristics of GS.

In our study, as in other published series, non-HH GS may occur both in left and right hemisphere epilepsies and have no lateralizing value (Tassinari et al., 1991; Tran et al., 2014).

Of note, we have also identified two patients with concordant lesional and VEEG data for ictal onset in parietal and occipital lobes. Only one published study describes the involvement of the parietal lobe in the genesis of GS (Kovac et al., 2015). In theseries of Tassarini et al. (1991), there was also one patient in whom GS was elicited by fusiform stimulation. In one of the patients (Patient 4), a left temporo-parietal dysembryoplastic neuroepithelial tumour (around the parietal operculum) was removed and the patient remained seizure-free, confirming this area as the epileptogenic region. In is not possible to exclude, however, propagation to more anterior temporal areas. In Patient 3, only scalp EEGs were performed and we have to acknowledge that this more posterior EEG activity may be due to propagation from fronto-temporal origins. The ictal EEG, however, shows rhythmic activity in bilateral posterior regions (figure 1B) with later propagation to the bilateral anterior lead (not shown). Furthermore, the concordance between MRI (figure 1A) and VEEG (both ictal [figure 1A] and interictal [figure 1C]), and the fact that these posterior brain regions have also been described to participate in the cerebral networks of laughter, suggest that these areas should also be considered as possible ictal onset zones for HH. Taken together, our data and the data from other series in the literature suggest that gelastic manifestation in seizures may not have localizing value.

Most patients with GS presented with refractory epilepsy, as is usually the case in GS series (Striano et al., 2009; Gutierrez et al., 2016). In this study, we aimed to focus on patients with GS in clinical practice, avoiding the reference bias associated with VEEG data. The frequent occurrence of refractory epilepsy in our series, however, may still be related to reference bias rather than a specific association with GS. On the one hand, we recruited patients from tertiary care hospitals with frequent referrals for refractory epilepsy. On the other hand, it is possible that subtle gelastic features of seizures in patients with other types of ictal semiology were missed in patients with complete seizure remission under medication or with rare seizures, as is the case of non-refractory patients. Of note, patients submitted to epilepsy surgery had an overall good surgical outcome, emphasizing the importance of this treatment strategy in selected patients (Téllez-Zenteno et al., 2010).

This study presents several limitations. First, it is retrospective and multicentric with possible selection bias, and included non-standardized patient evaluation. Secondly, the majority of the diagnosis of GS was made based on direct clinical observations or relatives’/caregivers’ descriptions. Although this may have led to false-positive cases, it also makes our sample more comprehensive because it also included patients for whom VEEG was not necessary. Finally, due to the retrospective nature of our study, other semiological data that may have clinical or localizing value, such as the emotional content of GS (laughter with no mirth originating in anterior cingulate areas (Sperli et al., 2006) and emotional content from more basal temporal areas (Tassinari et al., 1991; Fried et al., 1998), or the association with hormonal symptoms in HH, were not collected.

In conclusion, our study reinforces that, despite the strong association between GS and HH in children, other aetiologies and patterns of affected topography unrelated to HH are common in patients of all ages with these types of seizure manifestation. This ictal sign has no lateralizing value and, despite a clear preponderance for hypothalamic, frontal and temporal lobe origin, other brain areas, namely the parietal and occipital lobes, should be considered. Further studies are necessary to better characterize ictal laughter and its association with specific emotional content.

Supplementary data

Summary didactic slides are available on the www.epilepticdisorders.com website.

Acknowledgements and disclosures

The authors would like to thank CA Tassinari for very kindly sharing his 1991 work on gelastic seizures. None of the authors have any conflict of interest to declare.