Reversible hyperintensity lesion on diffusion-weighted MRI in hypoglycemic coma

Case report.

A woman aged 73 years with a history of insulin-dependent diabetes mellitus, controlled by oral medications, was admitted to our hospital after being found at home in an unresponsive state. The patient lived alone, and her son informed us that she had been well until the day before admission. On arrival the patient was in coma (Glasgow Coma Scale, 4; E1, V1, M2). The patient exhibited isocoria (3 mm/3 mm) and was reactive to light. Rolling eye movement, tetraparesis, and decerebrate rigidity to pain were evident. She had a regular heart rate, hypertension (blood pressure, 180/102 mm Hg), regular respiration, and normal temperature, and there was no hypoxia. On checking her status, we considered the possibility of basilar artery occlusion. We ordered diffusion-weighted MRI (figure 1) and MR angiography (MRA). MRI showed the presence of hyperintensity lesions within the bilateral internal capsule, corona radiata, and frontoparietal cortex, and MRA showed no major abnormality, including basilar artery occlusion. We subsequently checked blood glucose and blood gases, which showed that her blood glucose was 20 mg/dL. The patient was immediately administered 50 mL of a 50% glucose infusion. Her conscious level and tetraparesis improved soon after the glucose infusion, and she recovered completely after 10 days without neurologic deficit. Diffusion-weighted MRI after complete recovery (figure 2) showed regression of hyperintensity lesions within the bilateral internal capsule, corona radiata, and frontoparietal cortex. The patient’s hypoglycemia was due to low dietary intake and error of medications (she had taken a double dose of antidiabetic drugs). The duration of the patient’s hypoglycemia was not apparent.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Diffusion-weighted MRI on admission showing the hyperintensity lesions within the bilateral internal capsule, corona radiata, and frontoparietal cortex. Note that bilateral hippocampi do not disclose any hyperintensity lesions.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Diffusion-weighted MRI 10 days after glucose infusion showing regression of the hyperintensity lesions.

Discussion.

We report reversible hyperintensity lesions on diffusion-weighted MRI in humans with hypoglycemic coma. MRI in hypoglycemic coma mainly involves the cortex, internal capsule, basal ganglia, and hippocampus.3,8,9⇓⇓ The affected lesions in our patient were the bilateral internal capsule, corona radiata, and frontoparietal cortex. The basal ganglia and hippocampus were spared on diffusion-weighted MRI, whereas in severely affected patients, the basal ganglia is often involved. However, the involvement of basal ganglia is caused by hypoxia accompanied with hypoglycemia.4 We determined that our patient was in hypoglycemic coma without hypoxia. The patient’s respiration was regular and we performed blood gas analysis without supplemental oxygen, and confirmed that no hypoxia or respiratory depression was present. Furthermore, the sparing of the basal ganglia on diffusion-weighted MRI implied good results.

Several reports have been published on the affected lesions by hypoglycemia.5,6⇓ Hypoglycemia disrupts protein synthesis in the superficial layers of the cortex, in the caudate and putamen, and in certain cell populations of the hippocampus (CA1 and CA3 pyramidal cells and dentate crest granule cells). However, protein synthesis in the cerebellum and brain stem remains relatively unaffected because of the greater activity of the glucose transport mechanisms.5 After 30 minutes of hypoglycemia in rats, adenosine 5′-triphosphate levels are much higher in the thalamus (45% of control value) than in the cerebral cortex (23%), striatum (27%), and hippocampus (17%).7 Furthermore, the expression of receptors for excitatory amino acids is involved in the formation of damaged lesions by hypoglycemia. Excessive release of excitatory amino acids, particularly aspartate, results in selective neuronal necrosis in the neuronal injury caused by hypoglycemia.1,2⇓

Distinct from routine MRI, diffusion-weighted MRI can demonstrate an abnormal signal within minutes. Furthermore, compared with MRI, diffusion-weighted MRI defines only fresh lesions. The usefulness of diffusion-weighted MRI for patients with hypoglycemic coma has only been occasionally reported. The hyperintensity lesions on diffusion-weighted MRI are within the basal ganglia and temporal and occipital lobes 1 day after glucose infusion and within the bilateral basal ganglia, hippocampus, and cerebral cortex on day 5.8,9⇓ The patients in both reports never improved. There is no data reporting whether hyperintensity lesions on diffusion-weighted MRI in patients with hypoglycemic coma are reversible after glucose infusion.

In our report, we demonstrate in humans that the hyperintensity lesions on diffusion-weighted MRI are reversible. Furthermore, the patient recovered without any neurologic deficit, confirming that diffusion-weighted MRI is useful for predicting prognosis. If the hyperintensity lesions regress in the second image, the patient will likely recover. However, if the hyperintensity lesions do not regress in the second image, the outcome will be poor. We do not know the time necessary for hyperintensity lesions on diffusion-weighted MRI to disappear in humans after glucose infusion. In our patient, it occurred within 10 days.

Infarction is the most common cause of such a hyperintense lesion on diffusion-weighted MRI. In hypoglycemia, the incidence of cytotoxic edema, shrinkage of the extracellular space as a result of hypoglycemia, and failure of the ionic pumps of the cell membrane after energy depletion cause the hyperintense lesion on diffusion-weighted MRI. Distinct from infarction, intracellular acidosis does not occur.1 In hypoglycemia, tissue alkalosis occurs because of the formation of ammonia from the deamination of amino acids and the consumption of metabolic acids. Although infarction and hypoglycemia exhibit similar findings on diffusion-weighted MRI, their mechanisms are distinct.

Determination of the blood glucose levels and the effects of glucose infusion are useful tools for differential diagnosis. Another method is to compare the diffusion-weighted MRI on the day of admission to one taken several days after glucose infusion. In the case of infarction, a hyperintense lesion on diffusion-weighted MRI is unlikely to disappear within several days. In an animal study, severe transient hypoglycemia causes a reversible change in the apparent diffusion coefficient of water.10 The result in our patient is consistent with that study.