Our cases demonstrate 2 prolonged HM with delayed neuroimaging changes that both resolved after the acute period (). HM is a rare primary headache disorder normally characterized by less than 24 hours of neurological deficit.1
These cases are similar to several previously reported (),3–7
including prolonged attacks,3–7
or diffuse cortical edema,4
and occasional subtle diffusion abnormalities.4,6
All 7 cases had MRI changes and prolonged attacks. Although it has been estimated that 8% of SHM are prolonged,2
it does not appear that any particular gene is more likely to produce protracted attacks.
The mildly restricted diffusion seen in our cases and those in the literature likely corresponds to mild cytotoxic edema resulting from the HM attack.3
Although diffusion changes are commonly assumed to represent ischemic tissue, Chabriat et al propose that the subtle diffusion abnormalities in HM are the result of prolonged neuronal depolarization.8
Similarly, Kumar et al found no perfusion abnormalities on single-photon emission computed tomography (SPECT), but fluorine 18-labeled deoxyglucose (FDG) positron emission tomography (PET) demonstrated significant decrease in metabolic activity in the involved hemisphere, supporting the notion of neuronal depression.5
This could also explain why there is a lag between radiological and clinical findings that is not normally seen in ischemic events. The diffusion abnormalities in our cases were subtle, and the cortical swelling was not limited to any discrete vascular territory. The transient and late restricted diffusion changes may indicate metabolic stress where neuronal cells are unable to meet energy demands, resulting in dysfunction and mild swelling of cells without ultimate neuronal cell death.
Although most cases showed radiographic and clinical resolution, one case in the literature showed residual volume loss, and Case 1 in our series continued to have persistent neurological symptoms requiring an individualized learning plan. In addition, Dodick et al report an FHM case with acute hemispheric edema, no diffusion changes, and a persistent neurological deficit 2 years later.9
These cases suggest that permanent migrainous infarction occurs in a minority of patients. Hypometabolism on PET and reduced perfusion on SPECT 3 months after onset of attack suggests irreversible neuronal death without ischemia may also occur.9
Paradoxically, Herold et al have shown normal cerebral oxygen consumption during an episode of HM.10
Cerebellar atrophy can also be seen in adults with HM, which was not reported in any of the reviewed published pediatric cases or in our series. This could represent additional long-term sequela in HM.
Not all patients with HM are found to have a mutation in 1 of the 3 recognized genes; however, abnormalities in the identified calcium channel or Na+/K + pump may explain the changes thought to be related in HM. A possible explanation for neuronal compromise with prolonged attacks is neuronal excitability via increased ionic flux and glutamate and decreased glutamate uptake in HM.8,9
These changes may lead to prolonged cortical spreading depression and neuronal injury seen in HM.
In conclusion, the pathophysiology and significance of cytotoxic edema in childhood HM remains uncertain. Imaging changes may show mild hemispheric edema with or without diffusion changes, which are often subtle if present. Sporadic cases may be more commonly associated with ATP1A2 gene7
as opposed to CACNA1A. Prolonged HM may produce residual symptoms and persistent MRI findings, and it is unclear if there is underlying infarction vs non-ischemic neuronal cell death. Serial imaging during and after resolution of an attack may help elucidate the pathophysiology and spectrum of HM, including the use of functional imaging techniques such as PET and SPECT.