The four children, twin girls and two male siblings here reported, are representative example of typical AHC disorder. The presenting clinical manifestations of the children fell within the diagnostic criteria expressed by Krageloh and Aicardi (1980) [10] and Bourgeois et al. (1993) [11]. These diagnostic criteria are: a) onset before the age of 18-months; b) autonomic phenomena; c) cognitive impairment; d) repeated episodes of hemiplegia that sometimes involve both sides of the body; e) evidence of neurological abnormalities such as choreoathetosis; f) the symptoms disappear during sleeping and resume after waking up. The clinical signs of AHC are complex, heterogeneous, and follow unique a pattern: their clinical progression tends to occur in sequential distinctive phases, the paroxysmal episodes are often preceded by precipitating factors such as environmental stress, such as bathing and other events. Sleeping reliefs the symptoms. Non-paroxysmal features such as DD/ID, epilepsy, motor dysfunction, movement disorders and autonomic manifestations in association with the paroxysmal events are the cornerstone for making the diagnosis of AHC rather than being considered as comorbidities [12].
Mikati et al. (2000) have demonstrated that the clinical course of patients with AHC evolves in three distinct phases [12]. Phase one begins during the first few months of life and continues for one year and the most common features consist of unilateral nystagmus, ocular deviation, dystonic spells, and developmental delay. Phase two lasts from the age of one to five years, in which the hemiplegic spells become more typical, with a possible frequency of several times each month, and with a duration of several days or even weeks. In this phase, abnormal movements, dystonic attacks, and choreoathetosis are frequently observed. Phase three is represented by fixed neurologic deficits and obvious ID. In this phase, dystonic and hemiplegic episodes become less frequent and less severe. Main clinical manifestations compared to those indicated by Mikati et al. [12] are summarized in Table 1. Identification of this pattern and how the symptoms progress facilitated earlier diagnosis of this disorder bearing in mind that the symptoms are wide particularly regarding the duration and frequency of hemiplegic and dystonic episodes. The beneficial effect of sleep on abnormal paroxysmal features with the disappearance of paroxysmal phenomena and resumption of the normal movements is one of the diagnostic criteria of AHC. This feature is not always reported. However, remission of the paresis may be observed even after a short nap. A study on sleep architecture was carried out in four AHC children and the results showed a normality on the sleep structure, sleep duration, cycle length, rapid eye movement (REM) latency, and REM and slow wave sleep (SWS) percentages [13]. Clinical suspicion starts when the infant presents with abnormal ocular movements such as nystagmus and ocular deviation, head deviation, dystonic spells and unilateral hypotonia, which is usually triggered by several factors including light, sound, exposure to heat or cold, and stress whether physical or psychological. Paroxysmal hemiplegic episodes usually start after the first year of life and usually fluctuate from a side to the other or occur simultaneously on both sides. These episodes may be accompanied by speech impairment, gait incoordination, and movement disorders. The first diagnostic approach is to exclude a diagnosis of epileptic seizures, which can precede, co-occur with, or follow the hemiplegic attacks. A prolonged Video EEG is pivotal for differentiating epileptic seizures from the paroxysmal events of AHC. Routine laboratory examination, plasma amino acids, urine organic acids, blood lactate, pyruvate, urea, ammonia, thyroid functions, arterial blood gases (ABG), MRI and MRI angiography are effective to exclude metabolic disorders and vascular diseases having the same pattern of features such as homocystinuria, organic acidurias (glutaric aciduria), urea cycle disorders (ornithine transcarbamylase deficiency, carbamoyl phosphate synthetase I deficiency, and citrullinemia) and Moyamoya disease. Diagnostic check-up may also include analysis of pterins, 5-methyltetrahydrofolate (5-MTHF) and monoamine metabolites in the cerebrospinal fluid. Regarding the severity of the condition, AHC is usually reported as devastating since hemiplegic features are often associated to other neurologic dysfunctions including DD/ID and epileptic seizures, as discussed in the following paragraphs.
Mikati et al. (2000) [12] reported that developmental delay was observed in 40 out of 44 patients enrolled in their study. According to these authors, developmental level correlates with the age of AHC individuals and with the age of onset of the hemiplegic episodes. Although neuropsychological evaluation showed wide variability in functional impairment for cognitive, adaptive and behavioral domains, younger patients demonstrated better results. The authors suggest that remain to establish whether the cognitive delay in AHC individuals is related to the repeated attacks of hemiplegia or a primary effect of the disorder [12]. In the study of Sweney et al. (2009)[14], cognitive impairment was generally defined by families as mild to moderate, and recently a mild cognitive impairment form was reported by Polanowska et al. (2018) [15] in two adult patients, in whom a neuropsychological examination showed a normal or near normal global cognitive functioning, with only some isolated executive functions deficits. In the children here reported the ID was mild and without a progressive course.
Epileptic seizures are reported in about 50% of AHC individuals (Table 3) [12, 16–18]. In the study of Mikati et al. (2000) [12] only 8 (19%) out of 44 patients experienced epileptic seizures, which occurred infrequently in one-half of these patients (three seizures or less). Out of those eight patients, four patients presented with generalised tonic clonic seizures, three with focal clonic seizures, and one with generalised myoclonic seizures. Status epilepticus appeared only in one patient. According to Sweney et al. (2009)[14], 44 (43%) out 103 AHC individuals showed generalized tonic or tonic clonic seizures. The mean age of onset of epileptic seizures was around 6 years, with 10 (23%) of the 44 cases who did not experience epileptic episodes until the age of 10 years or later. Ictal seizure EEGs in AHC individuals was reported by Saito et al. (2010) [19]. In another study, status epilepticus appeared in 4 out of 9 patients at the age of 6–16 years [20]. In a report of Uchitel et al. (2019)[21] on 51 patients with AHC, 32 (62.7%) had focal epilepsy in different cerebral regions, but more frequently frontal region; 11 (21.5%) showed primary generalized seizures tonic clonic, myoclonic, and/or absence. In 8 (15.5%) patients, epileptic seizures preceded other AHC paroxysmal events. However, according to Heinzen et al. (2012) [5] epileptic seizures may precede the paroxysmal hemiplegic episodes and EEG registration may appear initially normal and then EEG may become epileptiform. In the present cases, the twins up to the age of 19 years never complained of seizures, whereas epileptic seizures were recorded in the siblings presenting with focal seizures with onset in one brother at the age of 3 ½ years, and in the other one, at the age of four years. EEG for both siblings showed multifocal spikes and waves expressed mainly in the frontal region. In general, epileptic seizures episodes are reported with low frequency and good response to treatment. A clinical distinction between episodes of hemiplegic attacks and epileptic seizures is not always clear, and the correlation between the epileptic and hemiplegic episodes remains doubtful [19]. Ictal episode was registered in one of the siblings here reported.
Migraine is a symptom not commonly found in AHC neither in the affected individuals nor in the family history [14]. In the twins here reported the episodes of migraine with aura was one of most consistent symptoms.
The cognitive impairment, epileptic seizures, persistent movement disorders, and autonomic dysfunction are considered comorbidities of the AHC disorder. Nevertheless, after refining the symptoms of AHC by Krägeloh et al.[10] in 1980, then by Bourgeois et al. [11] in 1993, it became obvious that these symptoms may be recorded as primary components of AHC. ATP1A3 has been implicated aside to AHC syndrome to other complex syndromes including the Rapid-onset Dystonia-Parkinsonism [6, 22, 23], and the Cerebellar ataxia, Areflexia, Pes cavus, Optic atrophy, and Sensoryneural hearing loss (CAPOS) syndrome [7, 24, 25]. At their present childhood/adolescent ages no one of the four cases here reported showed clinical features consistent with the uppermentioned syndromes.
Variability in clinical expression of paroxysmal and non-paroxysmal episodes in AHC individuals is well known. In the present cases the intrafamilial variable clinical expression was observed as regard to the course of the intensity and frequency of the hemiplegic episodes and of migraine attacks more pronounced in one of the twins (Table 1). The cognitive impairment was mild in both the twins and no seizures were recorded. In the siblings, the epileptic seizures were more severe in the child who showed more marked hemiplegic attacks. Cognitive impairment was mild in both siblings. It is presumable that epigenetic events have conditioned the intrafamilial clinical variability.
The diagnosis of AHC is mainly clinical but may be supported by molecular analysis. Typical gene mutations involved in the pathogenesis of AHC are located in ATP1A2 and ATP1A3 genes, but in some cases of AHC, however, recurrent mutations are not found. WES analysis has allowed us to explore the causative genes underlying the pathological mechanisms, which may contribute to the phenotypic variation observed between two siblings. We found in the oldest brother the identical haplotype inherited from the asymptomatic father, constituted of three heterozygous variants in GRIN2A (c.3175T > A), SCN1B (c.632G > A) and KCNQ2 (c.1870G > A) gene, while in the younger child, who has a milder phenotype, only the variant in GRIN2A gene. This mirrors the atypical pattern of inheritance of incomplete or reduced penetrance of AHC-2 mutations. The GRIN2A gene encoding the NMDA receptor (NMDAR) subunit GluN2A has been suggested to constitute a locus for mutations in a subset of individuals with early-onset seizures [26]. A literature overview shows that pathogenic or likely pathogenic variants are spread over nearly the entire gene, but the majority of null variants in healthy individuals are observed in the exon 14, which encodes nearly the complete C-terminal domain. These variants do not exhibit effects because this region represents the protein region tolerant to genetic variation in the general population. No missense variants in the intracellular C-terminal domain of GluN2A (beyond amino acid position 838) have been found to fulfil ACMG criteria for being pathogenic or likely pathogenic [26]. According to the pathogenicity prediction, the GRIN2A (c.3175T > A; p.Ser1059Thr) variant is likely pathogenic leading to significant alterations of the protein properties and splice site changes. It can be assumed as a first delineation of a phenotypically causal variant observed in the C-terminal domain of the protein associated with a reduced penetrance, since normal individuals are more likely to pass on their pathogenic variants to the offspring [26]. Additional two likely pathogenic variants implicated in childhood epilepsies [27–29] one of SCN1B (rs150721582) gene and another one of KCNQ2 (rs771211103) gene, were identified in the firstborn, who carried the more severe phenotype compared to the younger brother possibly due to harmful and cumulative effects driven by these mutations together with the GRIN2A variant. Some variant forms of AHC, AHC-related disorders and channelopathies are characterized by mutations involving glutamate, sodium and potassium channels [30–32]. The identification of mutations in GRIN2A, SCN1B and KCNQ2 genes shows the power of some low penetrant variants in eliciting the phenotypic spectrum of AHC not associated with the primary genetic risk factors in ATP1A2 and ATP1A3 genes [28, 32]. Analyzing the role of these genes both individually and in association with each other, we cannot exclude to be likely eligible as risk factors for AHC since the number and type of alteration in the expression of encoded proteins may disrupt some functional connections in the transcriptional regulatory network leading to the clinical variability of the disease.
To date, no drugs are available to cure AHC. The treatment usually comprises multiple drug therapy regimen. The aim of these therapeutic agents is prophylactic against the paroxysmal attacks. Flunarizine is a calcium channel blocker that has been widely indicated as the most effective drug for AHC treatment. This drug was first proposed by Casaer and Azou (1984)[33] in one child, and then in 12 children with AHC [34]. Subsequently to the first indication [33], the drug was tried by Sakuragawa (1992) on 23 patients [3] and by Burgeois and Aicardi (1993) [35] on 17 patients. The results obtained by these authors suggested that flunarizine therapy reduced the duration and severity of hemiplegic attacks, but did not interfere with the natural course of the disease. To note that no severe side-effects have been seen in any patients during the time of treatment [20]. In the study of Mikati et al. (2000) [12], 27 out of 44 patients affected by AHC showed a good response to Flunarizine. According to these authors [12], a clinically worthy reduction in the frequency and/or severity was found in 21 patients (78%). Flunarizine eliminated the hemiplegic attacks completely in one patient (4%), improved the attacks partially in two patients (7%), and was without any effect in four patients (15%). Sweney et al. (2009)[14] reported improvement in dystonic or hemiplegic episodes in 48 out of 80 patients using flunarizine and 21 out of 55 patients using benzodiazepines. In an Italian study conducted on 30 patients with AHC, Pisciotta et al. (2016)[6] reported that flunarizine resulted in reduction in the duration and frequency of hemiplegic attacks in 50% of the patients and decreased their intensity in 32.1%, the decreased intensity was more evident in the younger patients. No correlation between developmental outcome and duration of treatment was reported by these authors [6]. Flunarizine seems to be the best treatment option for the episodic hemiplegic attacks in patients with AHC. In our patients the treatment with flunarizine was conducted irregularly by the parents, however it was able to reduce frequency and intensity of hemiplegic attacks when regularly used. Recently, new drugs and new treatment modalities have been proposed. Triheptanoin is a triglyceride composed of three seven-carbon (C7:0) fatty acids. It acts by providing anaplerotic substrates for Krebs cycle and appears to increase the efficacy of the ketogenic diet. It was tested on 10 patients with AHC [20]. The trial with this drug was unsuccessful and failed to reduce the paroxysmal events. Aripripazole, an atypical antipsychotic drug was proposed by Dundar et al. (2019)[36] in a 12-year-old male with AHC who was not responding to any type of classical treatment; Aripripazole showed a reduction in the duration and frequency of hemiplegic episodes. Verapamil is a calcium- channel blocker and has the same mechanism of action of flunarizine. It was administered in a 5-year-old boy with AHC presenting with poorly controlled epileptic seizures and developmental delay [37]. Notable reduction in frequency, severity, and duration of the hemiplegic attacks was obtained with the use of this drug [38]. In association with flunarizine, anticonvulsant has been applied in the treatment of the epileptic attacks using benzodiazepine, carbamazepine, barbiturates and valproate [39]. In the siblings, lorazepam and benzodiazepines managed to control the seizures.
AHC is a complex, often serious disorder in which the hemiplegic episode is only one sign even if the most remarkable, of several other body-system impairments involving the autonomic nervous system, musculoskeletal system, and the brain, with epileptic seizures and cognitive dysfunction as a relevant clinical association. The narrow association and correlation of these clinical signs leads us to conclude that clinical involvement of AHC must be extended and included in the term “AHC spectrum disorder”.