This case presented with a variety of phenotypes that included fever, vomiting, alternating hemiplegia, convulsions, drowsiness, aphasia, episodic headache, unilateral cerebral oedema and cognitive impairment. HM caused by mutations in the ATP1A2 gene has been reported [6, 7]. However, it is rare to report so many symptoms that are recurrent and varied in a single child. In addition, the onset of the HM was atypical because no headache apeared at the onset. He was misdiagnosed many times.
In the early stages of the disease, fever, vomiting, drowsiness, and hemiplegia were observed. A head MRI showed unilateral parieto-occipital cortical oedema. The white blood cell count and protein, sugar and chloride levels in the CSF were normal. These manifestations were consistent with the diagnosis of encephalitis [8]. However, the patient subsequently had several similar episodes, and we believed that this patient might not be consistent with the diagnosis of encephalitis. At this time, because the patient presented with fever, vomiting, drowsiness, hemiplegia and convulsions, MRI showed unilateral cortical oedema, which occurred repeatedly. We considered that the child might suffer from mitochondrial encephalopathy induced by infection. However, the patient’s blood or cerebrospinal fluid lactic acid levels were normal many times, and there was no similar family history. Subsequent mitochondrial DNA sequencing analysis was normal. In general, the clinical data and related examinations did not meet the diagnostic criteria of mitochondrial encephalopathy.
Many of the patient’s clinical features supported the diagnosis of AHC, such as alternating hemiplegia, developmental retardation, and convulsions. In addition, NGS of the nuclear genome revealed a de novo heterozygous missense mutation in the ATP1A2 gene (p.Gly715Arg) classified as pathogenic. Mutations in the ATP1A2 gene can cause hemiplegic migraine type 2, common migraine, and basilar migraine but can also cause a very small number of AHC cases [9,10,11,12,13,14]. However, our patient did not experience improvement of hemiplegia during sleep, as described in previous reports [15, 16]. In addition, studies have shown that brain MRI of patients with AHC always shows normal or nonspecific findings [15, 16]. However, the brain MRI of our patient showed unilateral cortical oedema. In conclusion, the basis for the diagnosis of AHC in the child was still insufficient.
With the progression of the disease, the child developed paroxysmal headache, accompanied by hemiplegia and aphasia. Symptoms were recurrent and mostly reversible. First- or second-degree relatives had no migraine with aura including motor weakness. These core symptoms and characteristics were consistent with the diagnostic criteria of SHM [2]. It has been widely reported that HM can sometimes be accompanied by symptoms such as fever, convulsion, altered consciousness (drowsiness or even coma), and cognitive impairment, as well as cerebral oedema seen in head MRI [3, 6, 7]. All of these symptoms occurred in our patient. NGS of the nuclear genome revealed a de novo heterozygous missense mutationin the ATP1A2 gene (p.Gly715Arg) classified as pathogenic and eventually helped him to be diagnosed with SHM when he was 11 years old. The duration of symptoms is usually 20–60 min. In some cases, the aura and hemiplegia may onset quickly and simulate an ischaemic attack [17]. Complete recovery from attacks is the rule, but in severe HM, hemiplegia and altered consciousness may persist for weeks until normal [3]. In this case, symptoms lasted from 3 days to 1 week until total recovery. The child had recurrent headache attacks for approximately 2 months before the fourth attack, and headaches became more intense 1 week before admission. As the child grew older, the headache increasingly affected his life. In addition to this, he only had four episodes. However, as the disease progressed, the patient developed cognitive impairment, and cognitive decline was a serious problem. Consequently, flunarizine was subsequently administered, and it has now lasted for 4 months without any adverse effects. No recurrence was found during follow-up. However, the exact length of treatment is still unclear because this is an atypical case, and follow-up of his symptoms and regular monitoring of the side effects of flunarizine will help determine this. In the case report, we found a significant phenomenon. In addition to the evolution of symptoms, the child’s MRI evolved from cerebral oedema at the beginning of the disease to normal MRI at the last attack. However, the picture of the MRI has been lost. This is the limitation of this case report, but it did not affect the conclusions because we knew the MRI findings from the MRI report provided by the guardian. In addition, when the third episode occurred, we performed an EEG on the child when they had hemiplegia but did not complain of headache. The EEG showed diffuse slow waves and no epileptic discharge. However, it was regrettable that we did not perform long-term EEG in this child to understand the ictal and interictal EEG manifestations. This is conducive to our recognition of HM [18]. At present, the pathogenesis of migraine is still unknown and focuses mainly on neurophysiology, especially cortical brain disorders [19,20,21]. This needs further study in the future.
In summary, HM in children may be atypical in the initial stage of the disease, which could manifest as fever, alternating hemiplegia, and drowsiness but no headache at the onset. This could easily lead to misdiagnosis. With age, it may eventually manifest as typical HM. Symptoms were recurrent and variable. Therefore, attention should be given to differentiation in clinical practice. When similar clinical cases appear, gene detection is particularly important, which is conducive to early diagnosis and treatment.