Transient lesion in the splenium of the corpus callosum in an epileptic patient

Case report.

A 28-year-old, left-handed woman developed intractable cryptogenic epilepsy with simple or complex partial seizures (CPS) and rare secondary generalization at age 13 years. Two or three clusters of CPS per month occurred with right hemifield defects, epigastric sensation, and fear, followed by impaired awareness with oroalimentary, bimanual, bipedal, and truncal automatisms. Postictal dysphasia occurred during some episodes. The simple partial seizures consisting of episodes of flashing small white and colored circular lights in the right superior visual quadrant, with right hemifield amaurosis and rightward head and eye deviation did not occur until age 26.

Daily antiepileptic drugs (AEDs) on admission to our epilepsy monitoring unit included 1600 mg of carbamazepine (serum level, 25.4 μmol/L; normal range, 17 to 51) and 400 mg of lamotrigine. She had been prescribed carbamazepine for 15 years and lamotrigine for 2 years. Her medical and family histories were normal. Physical examination, including visual fields and detailed neuropsychological testing, was normal.

Scalp EEGs were normal except for minimal theta in the left posterior temporal region and rare left occipital interictal spikes activated by sleep. Seizures originated in the left occipital or occipital-posterior temporal regions with secondary generalization. Three previous 1.5-T cranial MRIs and two cerebral angiograms obtained during intracarotid amobarbital tests at a local institution were normal. Her last MRI was obtained 7 months before her admission to our epilepsy center.

During the first 6 days of the telemetry monitoring, carbamazepine was reduced to one-half and lamotrigine was reduced to two-thirds of the admission doses. She received one-third of the carbamazepine dose and two-thirds of the lamotrigine dose for an additional day. Carbamazepine was then discontinued for 5 days, and lamotrigine was maintained at one-half of the original dose. She received one-third of the original carbamazepine dose and two-thirds of the lamotrigine dose the next day. Five focal seizures with secondary generalization occurred in less than 48 hours, 4 days after carbamazepine omission. Our standard 1.5-T MRI to exclude cortical dysplasia of the occipital lobes, obtained 16 hours after her last seizure, revealed a homogeneous circular lesion 1 cm in diameter with distinct margins in the splenium of the corpus callosum (SCC; figure 1). Diffusion-weighted imaging (DWI), fluid-attenuated inversion recovery (FLAIR), and T2-weighted images disclosed an increased signal abnormality, and T1-weighted images showed a slightly reduced signal abnormality that did not enhance after gadolinium contrast injection. The apparent diffusion coefficient (ADC) of the SCC lesion was 0.15 × 10⁻³ mm²/s, much lower than the ADC values of the right frontal white matter (0.76 × 10⁻³ mm²/s), right frontal gray matter (0.81 × 10⁻³ mm²/s), and CSF (1.34 × 10⁻³ mm²/s). There was no detectable abnormality on the anatomic images of our 4-T fMRI obtained 1 day before her seizures, i.e., 5 days before 1.5-T images.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. T2-weighted axial (A), coronal fluid-attenuated inversion recovery (B), and diffusion-weighted trace axial (C) images showing abnormal signal intensity in the splenium of the corpus callosum.

Neurologic examination remained normal with no evidence of disconnection syndrome. Her neuropsychological examination had been completed 1 day before her abnormal MRI findings. Repeated neuropsychological studies pertaining to visual cortex, parietal lobes, speech, language, reading, writing, and praxis failed to show any new findings. Imaging features, very low ADC1 in DWI in particular, and ancillary tests rendered other causes, including MS, unlikely. Goldmann visual field testing, fMRI studies of the visual system, and auditory and somatosensory evoked potentials were all normal. CSF studies and blood work were unremarkable for a demyelinating illness, connective tissue diseases, autoimmune or other systemic conditions, infection, metachromatic leukodystrophy, or antiphospholipid antibody syndrome. Other causes of a focal splenial lesion, such as hemolytic-uremic syndrome, HIV infection, trauma, Marchiafava–Bignami disease, and acute mountain sickness, were believed to be unlikely. The amplitudes of evoked responses to patterned visual stimulation were lower to the right hemifield stimulation, possibly reflecting cortical dysfunction related to the left occipital seizures.

To avoid possible MS exacerbations, the patient’s investigations for epilepsy surgery were postponed 3 months. Complete resolution of the splenial lesion on repeat DWI, T2-weighted, and FLAIR images 3 months later rendered infarct unlikely (figure 2). The ADC of SCC increased to 0.84 × 10⁻³ mm²/s, comparable with ADC values of the right (0.87 × 10⁻³ mm²/s) and left (0.83 × 10⁻³ mm²/s) frontal white matter.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. T2-weighted axial (A), coronal fluid-attenuated inversion recovery (B), and diffusion-weighted trace axial (C) images showing complete resolution of the abnormal signal change seen in figure 1.

Three weeks of subdural recordings with lines over the occipital, posterior parietal, and temporal lobes revealed left mesial occipital origin of spikes and seizures with prominent spread to the temporal regions bilaterally. All 10 stereotyped clinical and electrographic seizures originated from the left mesial occipital region and spread prominently to both temporal lobes; the minimal homotopic involvement suggested, at most, slight propagation via the SCC (figure 3). Six of the seizures were recorded over 5 days, 1 week before her postoperative MRI. A 200-mg reduction in carbamazepine and a 25-mg reduction in lamotrigine daily doses during the second monitoring week were the only AED changes. Postoperative MRI did not detect any new lesions. No histologic diagnosis of the resected tissue was available.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. A stereotyped subdural-recorded seizure 3 months after MRI finding of the splenial lesion. It shows onset of the seizure in the left mesial occipital region (70s) with propagation to the temporal lobes. Only involved channels are displayed: 10s (left mesial temporal), 20s (right mesial temporal).

Discussion.

MRI features of high signal abnormality on T2-weighted, DWI, and FLAIR images with low signal abnormality on T1-weighted MRI sequences and complete resolution in 3 months likely represent transient ischemia and cytotoxic edema from frequent seizures, rapid carbamazepine reduction, or a combination of both.1,2⇓ Normal cerebral angiograms and MRI studies excluded an underlying vascular abnormality as a predisposing factor. Lack of ictal spread through the SCC in the subdural studies 3 months later provided some evidence that ictal propagation was less likely but still a possible cause of the splenial lesion. No clinical or laboratory data supported other causes of a focal SCC lesion, including MS or nutritional deficiency.

Similar transient lesions of the SCC have previously been reported in 16 epilepsy patients. They were attributed to focal edema that originated from transcallosal propagation of seizures in seven patients with temporal lobe epilepsy (TLE) monitored in an epilepsy unit. Five of these seven patients had secondary generalized seizures.3 Review of MRIs of 1,200 epileptic patients revealed such transient lesions in six patients with CPS.4 Seizure origins were not specified, but the patients likely had TLE because four had mesial temporal sclerosis. The lesions completely resolved after discontinuation of phenytoin and vigabatrin in two of the patients. Therefore, they attributed the lesions to AED-related demyelination. Another group reported such lesions after abrupt fluctuations of AED levels in three monitored patients.5 They attributed the lesions to the effect of AEDs on arginine vasopressin (AVP) hormone and fluid balance.5

Frequent seizures, AED fluctuation with a consequent impact on AVP and tissue fluid balance, or a combination of both produced the transient lesion in our patient. Seizure-related MRI changes have been previously reported.6 AVP regulates regional cerebral blood flow7,8⇓ and brain water content.9 An adaptive phenomenon, serum AVP level decreases after several weeks of carbamazepine therapy.7 With continuous therapy, the fluid balance system adapts to the influences of the AEDs. Abrupt serum carbamazepine concentration decrease may elevate AVP, destabilizing brain water balance. This could produce cytotoxic edema in vulnerable regions such as the SCC, which is solely supplied by the terminal branches of the posterior cerebral arteries.10

Frequent seizures likely played a role in some of the previously reported cases because frequent seizures were reported in all seven monitored patients in one study3 and in another monitored patient who experienced six seizures within 24 hours 4 days before his MRI-detected splenial lesion.5 Conversely, no apparent seizures were present in two other monitored patients in the same epilepsy unit,5 and none was reported within 3 to 4 weeks of the MRIs in any of the six outpatients in another study.4 AEDs and their rapid withdrawal or toxic levels were presumed to be the cause of the splenial lesions in the latter cases. Therefore, splenial lesion may affect a heterogeneous population of patients with epilepsy. It may require one of the aforementioned two elements in some patients and both in other patients. Our patient demonstrated presence of both factors; therefore, it is difficult to strongly favor one factor over the other. This may partly explain the rarity of this abnormal MRI finding among patients with epilepsy.

Our MRI and EEG data suggest that rapid AED reduction could have caused AVP levels to increase abruptly, producing transient focal ischemia evident by cytotoxic edema. However, we cannot exclude the effect of frequent seizures. Future MRS studies may help to differentiate seizure-related signal abnormalities on MRI from the effect of AEDs.