We read with interest the report by Lefebvre et al,1 of a case of severe bilateral visual loss after anaphylactic shock associated with isolated, symmetric, bilateral lateral geniculate nucleus (LGN) lesions on MR imaging studies. They concluded their patient’s visual loss was caused by isolated ischemia of the LGN from shock-induced hypotension. They stated the LGN can be considered part of the watershed region of the brain because of its proximity to this vascular border zone, even though ischemic damage to the LGN from prolonged hypoxia had never been reported.
We believe isolated extrapontine myelinolysis, or osmotic demyelination, is a more plausible explanation for their patient’s symmetric bilateral LGN lesions and visual loss. Central and extrapontine myelinolysis has been reported after resuscitation for anaphylactic shock.2
In 1987, Gocht and Colmant3 reported their findings in 58 autopsied cases of central and extrapontine myelinolysis, and found the LGN was the second-most-common extrapontine site of myelinolysis, occurring in 14 of 47 nuclei they examined. They reported that extrapontine lesions could occur in the absence of central pontine lesions.
In his review of the osmotic demyelinating disorders, Brown4 noted that central and extrapontine myelinolysis occurs in at-risk patients who are given intravenous fluids for treatment of an underlying disease. Osmotic demyelination can occur even when serum sodium levels are normal.
One of us (R.K.I.) reported a patient with presumed intrageniculate myelinolysis after near-fatal uterine hemorrhage and hysterectomy5 whose MR imaging appearance and visual field findings were strikingly similar to the patient reported by Lefebvre et al.1 Even though that patient’s serum sodium levels remained within normal limits, rapid shifts in calculated serum osmolality were demonstrated.
The acute demyelinating lesions seen on MR imaging studies in patients with osmotic demyelination are symmetric, hypointense on T1-weighted images, hyperintense on T2-weighted images (T2WI) and on fluid-attenuated inversion recovery images, and enhance with gadolinium. We believe the small hypointense areas seen on the coronal gradient echo T2WI in the patient reported by Lefebvre et al1 are consistent with acute necrosis but do not indicate cause.
Originally a neuropathologic diagnosis, central and extrapontine myelinolysis have increasingly become neuroradiologic diagnoses, because most patients survive and some even recover completely. Without neuropathologic confirmation, we can only speculate about the cause of these patients’ spectacular LGN lesions and visual loss. We believe, however, that the evidence favors extrapontine myelinolysis, not ischemia.
Reply:
We read with interest the letter by Imes and Barakos. The case Imes reports1 is indeed strikingly similar to ours; however, although we can only speculate about it, we conclude that ischemia caused the lateral geniculate nuclei (LGN) lesions in our case.
Visual impairment has indeed been noticed within the first 24 hours after shock. Onset of symptoms during central pontine myelinolysis (CPM) is typically delayed from 2 to 7 days.1 Our patient was in otherwise good health and did not have underlying conditions2–4 such as alcoholism, malnutrition, liver or kidney deficiency, neurologic disease, or malignancy, which are commonly seen in case of CPM or extrapontine myelinolysis (EPM). She developed no neurologic symptom pointing to a pons lesion, and neurologic examination was unremarkable. Serum sodium level remained within the normal range, at the emergency department (143 mEq/L) and at days 1 and 5 after shock (140 and 138 mEq/L, respectively). Shock resolved promptly with intravenous adrenaline and methylprednisolone. No massive intravenous fluid was administered. MR imaging demonstrated lesions located exclusively within both LGN. The similarity between our case and that of Imes et al1 is the hyperintense aspect on T2-weighted images (T2WIs), but the major difference is the evolutionary aspect on T1-weighted images (T1WIs). This latter aspect on T1WIs, but also on T2WIs, points toward an ischemic origin with a hemorrhagic component.5,6 Moreover, lesions enhanced after gadolinium (Gd) injection on the images obtained at day 6, which is again a normal finding in case of ischemia.5,6 In viewing the T1WI at first (day 1), no abnormalities were seen. In contrast, on day 16 lesions were hyperintense on T1WI and hypointense on gradient echo T2WIs and, thus, strongly suggest the hemorrhagic nature of the lesions in our patient. On day 51, lesions were still hyperintense on T1WIs, and hypointensity remained on gradient echo T2WIs, which is still in accordance with the findings reported in cases of hemorrhagic ischemia.5,6 Twelve months later, T1WIs returned to normal, as well as T2WIs and fluid attenuated inversion recovery images, but hemosiderin deposits were still perceptible thanks to gradient echo T2WI. The evolutionary aspect of our lesions, as seen on MR images, are in complete concordance with what is reported on MR imaging in cases of hemorrhagic infarct.5,6 To the best of our knowledge, this hemorrhagic aspect has never been described in myelinolysis.4 The clinical features and the typical aspect on MR imaging leads us to conclude still that the possible origin in the case we reported is ischemia rather than myelinolysis. In contrast, hypointense areas are seen on T1WIs in cases of acute myelinolysis, but nonhemorrhagic ischemic lesions behave the same way.5,6 Table 1 states the evolutionary aspects of the MR images in our case.
Moreover, the hypothesis of ischemia is supported by an experimental study on rats by Bauman et al7 They demonstrated that fluid percussion injury cause neuronal death in several thalamic nuclei, which are located within watershed areas, and that additional hypoxemia worsens cell loss in the dorsal LGN. The dorsal part of LGN seems to be more vulnerable to hypoperfusion because blood vessels enter the ventral surface and then run vertically through the LGN to end.8
To the best of our knowledge, there is no report on CPM affecting exclusively the LGN, except the case of Imes et al. Stanescu et al3 describe hyperintense lesions on T2WIs within the brain stem, the thalami, and the cerebellum after an anaphylactic shock. They do not mention the aspect on T1WIs, and lesions are not located within the LGN. Gocht and Colmant4 examined 58 brains with CPM. LGN myelinolysis was present in only 4/58 (7%), and their report does not mention whether these lesions were isolated or associated with EPM elsewhere. The authors they refer to never faced isolated LGN myelinolysis.
The article by Imes et al1 is particularly interesting to us because of the striking similarities with ours in terms of vascular collapse, visual field defects, and isolated lesions within both LGN. Their case, however, seems to differ from ours in terms of delay for symptoms and recovery, massive intravenous fluids given, and the evolutionary aspects of the lesion on MR imaging. Ischemia in our case seems the most appropriate causative agent having led to LGN dysfunction.
Evolution of the signal intensity within the LGN on MR images
- Copyright © American Society of Neuroradiology
