Clinical manifestations

Presentation and course

Epileptic spasms (hereafter, spasms) consist of brief contractions of the axial musculature that may be in flexion, extension, or a combination thereof, typically accompanied by a concomitant extension of the upper extremities, and often brief but prominent opening of the eyes (35; 44). Individual spasms last a fraction of a second (usually 0.5 to 1.0 second) but characteristically recur in clusters, with repetitions approximately every 5 to 45 seconds, with individual clusters usually lasting several minutes and rarely more than 1 hour (30). Clusters are most frequent upon awakening in the morning but may occur at any time of day, and multiple clusters in a single day are not uncommon. There is great heterogeneity in semiology, and spasms may be symmetric or asymmetric, with lateral deviation of the eyes or head, or involving only 1 side of the body. Individual spasms are often followed by brief (ie, 5 to 10 seconds) spells of crying, which may reflect emotional and limbic phenomena; we have observed several cases in older children who report feelings of tremendous fear intermixed in a cluster of spasms. In many cases, spasms may be extraordinarily subtle with only minimal axial or appendicular movement, and perhaps only a modest but stereotyped change in facial expression or eye opening. Identification can be exceptionally challenging in these subtle cases, especially for the practitioner who has not actually seen the events and must depend on caregiver report. Diagnosis and treatment are often substantially delayed and inadequate awareness of spasms is at least partly responsible for preventable and potentially significant delays (19). The contemporary use of video documentation by parents has greatly aided identification.

The onset of spasms is before 1 year of age in the vast majority of cases, with peak incidence between 3 and 7 months (30). However, spasms may begin at birth, or they may appear long after the age of 12 months (33), with rare reports of incidence in adolescence (01). In cases with prompt identification and successful treatment, intellectual and epilepsy-related outcomes can be excellent (24; 10). However, given the challenges of identification, potential delays in effective treatment, and frequent underlying causes of spasms that are associated with poor intellectual outcomes independent of seizures (eg, trisomy 21), intellectual outcomes are often poor (35; 29).

Prognosis and complications

The outcome of spasms is generally poor but highly variable (38; 10). Key factors include whether or not treatment is prompt and successful, and whether identified causes of spasms are associated with poor epilepsy and developmental outcomes regardless of the presence or absence of spasms (eg, nonketotic hyperglycinemia) (29). In cases in which there is no developmental impairment prior to the onset of spasms and no identifiable cause of spasms despite a thorough work-up (classically termed cryptogenic), outcomes may be excellent (normal intellect and seizure freedom) with prompt and successful therapy.

Clinical vignette

For the past 4 weeks, a 7-month-old boy has had stereotyped episodes manifesting with a sudden extension of the neck and tonic stiffening of the bilateral upper extremities, occurring in clusters each morning. Gastroesophageal reflux was initially suspected, but empiric treatment with an antihistamine, and subsequently, proton-pump inhibitor was unsuccessful. Since the onset of the episodes, the boy has not been able to sit independently and has seemed less interactive. The developmental regression prompted neurologic consultation, which led to video EEG monitoring that then demonstrated epileptic spasms. On physical exam, several hypopigmented ellipsoid macules were seen on the trunk and extremities. MRI brain demonstrated subependymal nodules and scattered cortical tubers. A presumptive clinical diagnosis of tuberous sclerosis complex was later confirmed by genetic testing.

Biological basis

Localization

The localization of spasms is highly varied, and they may appear focal or generalized from the standpoint of clinical semiology, as well as ictal electroencephalography. Indeed, we have observed many cases due to focal cortical dysplasia (FCD), with sustained seizure freedom following resection of cortical dysplasia, in which no clues of focality could be discerned from semiology on video review or ictal electrographic signature. In these cases, focality was established with MRI, interictal positron emission tomography (PET) (09), and interictal intraoperative electrocorticography. In cases of focal and generalized origin, widespread cortical dysfunction is often demonstrated by (1) hypsarrhythmia and related patterns on interictal EEG, (2) widespread cortical hypometabolism on interictal brain PET, and (3) often simultaneous relative hypermetabolism in the caudate nuclei and other deep brain structures (08; 07). Moreover, this widespread cortical dysfunction, even in the face of focal sources of spasms, is thought to be responsible for the cognitive and developmental regression that frequently accompanies the early progression of spasms.

Pathophysiology

A precise pathophysiologic characterization of spasms has proven elusive. Though great progress has been made in the last decade, a new crop of animal models leave much to be desired as they do not adequately mirror the phenomena in humans; no single model based on a known cause of spasms in humans has been shown to exhibit spasms with ictal electrodecremental response, hypsarrhythmia, enduring unprovoked seizures, analogous neurodevelopmental sequelae, and responsiveness to the therapies that have been proven effective in humans (14). It is perplexing that so many disparate structural, metabolic, and genetic disorders have been associated with spasms, and yet among patients with these disorders, those manifesting comorbid spasms represent a small minority. The presumably genetic factors that distinguish patients at risk for spasms from those who actually manifest spasms are unknown. Such a discovery could facilitate the design of an ideal animal model and ultimately lead to superior therapies.

Epidemiology

The incidence of spasms is approximately 1 in 2500 live births (31; 13). Risk factors for the development of spasms include almost any neurologic insult, though the most common etiologies include hypoxic ischemic encephalopathy, malformations of cortical development (eg, focal cortical dysplasia), and tuberous sclerosis complex.

Prevention

There is no established strategy to prevent spasms. However, more data have been generated that spasms could be prevented in select at-risk populations, such as infants with tuberous sclerosis complex. A single retrospective study suggested that premorbid treatment of infants with tuberous sclerosis might effectively prevent infantile spasms and improve developmental outcomes (23). This led to a European-based multi-institutional trial to test the efficacy of vigabatrin on children with tuberous sclerosis before the development of epilepsy. Preventative vigabatrin use significantly delayed the onset of clinical seizures and the development of medication-resistant epilepsy (26). A similar clinical trial (PREVeNT Trial) has been undergoing in the United States to test the efficacy of preventative use of vigabatrin on development and seizure outcomes in infants with tuberous sclerosis (NCT02849457).

Differential diagnosis

Confusing conditions

The identification of spasms from clinical observation alone can be exceptionally challenging (27). Accurate diagnosis requires both ictal and interictal video-EEG monitoring (30). The most common mimics of spasms include Sandifer syndrome, fairly stereotyped axial contortions associated with infantile gastroesophageal reflux, myoclonic seizures and sleep myoclonus, brief tonic seizures, and a wide array of rapid stereotyped behaviors that are normally, if not commonly, observed in healthy infants. Rapid and correct differential diagnosis of suspicious events is essential as treatments for the above entities are radically different than spasms, and the neurodevelopmental cost of incorrect diagnosis and delayed treatment can be enormous.

Associated and underlying disorders

Spasms are a key element of West syndrome (the triad of spasms, hypsarrhythmia, and intellectual disability) and Ohtahara syndrome (early infantile epileptic encephalopathy), and they occasionally persist in children who develop Lennox-Gastaut syndrome. More generally, spasms occur as a consequence of myriad diseases and syndromes that affect the central nervous system in infancy. Although spasms may commence in older children and adults, spasms are highly age-specific with the vast majority of cases presenting in infancy and resolving by the 2 to 4 years of age. As it pertains to etiology, spasms are highly nonspecific. The most common associations with spasms include focal cortical dysplasia, hypoxic ischemic encephalopathy, tuberous sclerosis complex, stroke, intracranial hemorrhage, hemimegalencephaly, lissencephaly, polymicrogyria, trisomy 21, Aicardi syndrome, nonketotic hyperglycinemia, Menke disease, neurofibromatosis type 1, mutations of the ARX, CDKL5, and ASTXBP1 genes, and deletions and duplications of the chromosome 15q locus. Contemporary associations with numerous other gene defects are rapidly being identified with the expanded use of whole exome and whole genome sequencing algorithms.

Diagnostic workup

The cornerstone of diagnosis is video-EEG characterization of both events as well as the interictal background. Ictal EEG allows the distinction of spasms from normal behaviors and other types of seizures, and EEG characterization of sleep permits an adequate characterization of interictal patterns associated with spasms, namely hypsarrhythmia and related patterns. Hypsarrhythmia—the interictal phenomena defined by exceptionally high voltage (non-epileptiform slow-waves greater than 200 µV), disorganization, and multifocal epileptiform discharges (15; 18)—is thought to be a key marker for both the widespread cortical dysfunction neurodevelopmental sequelae, but its identification can be rather challenging (22).

Once spasms are confirmed, a search for the underlying etiology of spasms is critical, as this may impact the choice of initial therapy. Informed by clinical history, typical diagnostic approaches usually begin with brain MRI to identify structural causes of spasms (eg, focal cortical dysplasia, cortical tubers, hemimegalencephaly, neoplasms, stroke, and other vascular phenomena). In cases without identified structural abnormalities, the diagnostic approach then focuses on a search for genetic and metabolic entities such as Down syndrome, nonketotic hyperglycinemia, pyridoxine deficiency, and countless others. Contemporary approaches to genetic testing include targeted gene testing in cases in which a specific genetic disorder is suspected. In cases with high pretest suspicion of a genetic cause, but low pretest suspicion of a specific gene defect, contemporary approaches typically include chromosomal microarray analysis, epilepsy gene panels, and whole exome/genome sequencing (45).

Management

Spasms are often refractory to traditional antiseizure drugs and may be exacerbated by agents that inhibit voltage-gated sodium channel currents (eg, carbamazepine, oxcarbazepine) (39). Among commonly prescribed antiseizure drugs, topiramate (16; 41), zonisamide (37), and valproate (06), appear to exhibit only modest efficacy.

The most favorable response rates have been reported with the use of hormonal therapies including natural and synthetic adrenal corticotropic hormone (ACTH) (02) and corticosteroids including prednisolone (21). However, these therapies are associated with myriad side effects—the most important of which include immunosuppression, hypertension, and hyperglycemia. Although there is debate as to the ideal agent and dosage of hormonal therapies, and few head-to-head comparisons of competing regimens, high-dose protocols seem to confer additional benefit. At our center, we advocate for high-dose prednisolone (8 mg/kg/day, max 60 mg/day, divided in 3 daily doses) and high-dose ACTH (150 IU/m2 body surface area/day, divided in twice daily intramuscular injections) (21).

Vigabatrin is considered first-line treatment at many centers, especially for patients with tuberous sclerosis complex (40). Like hormonal therapy, the highest response rates have accompanied the use of relatively high dosage (100 to 150 mg/kg/day divided in 2 daily doses) (12). The use of vigabatrin is limited by fears of irreversible peripheral vision loss (11) as well as rare but potentially fatal phenomena associated with subcortical MRI signal changes (17). However, some of these fears may be overstated. At least with respect to vision loss, the incidence of visual field constriction (32) and electroretinographically-defined retinotoxicity (43) appear to be low with short courses of vigabatrin, and the prevalence of clinically-apparent vision loss is nearly zero (34).

Of note, a randomized controlled trial demonstrated that the combination of vigabatrin and hormonal therapy (high dose prednisolone or synthetic ACTH) resulted in higher short-term response rates and more rapid resolution of spasms (28), and early clinical response to treatment was associated with improved developmental and epilepsy outcomes at 18 months (28).

Though supportive data are less compelling, a variety of nonpharmacologic therapies have been advocated for use in treatment of spasms, including ketogenic diet therapy (25; 20) and vagal nerve stimulation (05).

Among contemporary developments, great fanfare has accompanied the use of artisanal cannabis extracts in the treatment of several severe forms of childhood epilepsy, including epileptic spasms (22). Several randomized controlled trials have been evaluating cannabidiol's efficacy, safety, and tolerability (NCT02953548; NCT03421496). Given the relatively fast progression of cognitive impairment that typically accompanies spasms and hypsarrhythmia, treatment is a relative emergency and should ideally begin within days of spasms onset and immediately on diagnostic confirmation. Similarly, response to a given therapy must be accomplished quickly and confirmed with extended video-EEG to ensure that neither hypsarrhythmia nor subtle spasms continue. Most contemporary treatment protocols require response within 2 weeks and immediately transition nonresponders to alternative agents. As an example, the Canadian Pediatric Epilepsy Network has demonstrated success with a protocol that begins with vigabatrin, with nonresponders transitioning to synthetic ACTH at the 2-week time point, and then transitioning patients who fail both vigabatrin and ACTH to topiramate at the 4-week time point (04). There are no adequate guides to therapy for patients who have failed first-line agents.

Outcomes

The long-term outcome of spasms is highly variable with respect to both epilepsy outcomes and neurodevelopmental outcomes. In considering infants (ie, IESS, including West syndrome), a significant minority become seizure-free and developmentally normal by age 2 years (13). However, mortality rates are markedly elevated throughout life and the leading causes of death are pneumonia among infants and sudden unexpected death in epilepsy (SUDEP) in adults (36). Most survivors who do not attain seizure freedom by age 2 years transition to other forms of epilepsy (eg, Lennox-Gastaut syndrome) and exhibit developmental impairment.