Situation (hyperglycemia)-related non-convulsive status epilepticus detected by 1.5-Tesla arterial spin labeling perfusion imaging in a patient with newly diagnosed diabetes during a neurological emergency

Akifumi Yokomizo; Takato Morioka; Fumihito Mugita; Satoshi Inoha; Tomoaki Akiyama; Takafumi Shimogawa; Nobutaka Mukae; Ayumi Sakata; Hiroshi Shigeto; Koji Yoshimoto

doi:10.3805/eands.A000167

Abstract

Situation-related non-convulsive status epilepticus (SR-NCSE) is a type of NCSE that is associated with various situations, including hyperglycemia. A 67-year-old woman with no prior history of type 2 diabetes mellitus visited our emergency room, experiencing difficulty typing on her mobile phone for approximately 2 h. Her blood glucose level was 334 mg/dL. 1.5-Tesla arterial spin labeling (ASL) perfusion magnetic resonance imaging, performed 40 min after her arrival, revealed presumed ictal hyperperfusion in the left parietal lobe. SR-NCSE associated with hyperglycemia was highly suspected, and she was immediately administered with insulin and antiseizure medications. The patient’s symptoms improved promptly. An electroencephalogram performed approximately 2 days later failed to reveal paroxysmal discharges. During neurological emergencies, ASL has the advantage of taking less time to perform than EEG, and we believe that ASL is useful in diagnosing SR–NCSE as well as structural focal NCSE when EEG cannot be performed.

Introduction

Situation-related non-convulsive status epilepticus (SR-NCSE) was first reported in 1992 by Thomas et al.[1] as a ‘de novo’ absence status of late onset in 11 cases. As the name implies, SR-NCSE is a type of NCSE caused by various situations, such as benzodiazepine-based drug withdrawal, metabolic factors including hyponatremia and non-ketotic hyperglycemia, or a combination of the forementioned[1,2,3,4,5,6]. Traditionally, electroencephalography (EEG) is essential in diagnosing SR-NCSE[1,2,3]. Recently, long-term video-EEG monitoring has been recommended for the accurate diagnosis of NCSE in Europe and the United States[7]. However, in Japan, many acute care hospitals do not possess long-term video-EEG monitoring facilities. Even routine EEG cannot be performed outside of consultation hours. Hence, the diagnosis and subsequent treatment are often delayed[4, 8,9,10,11,12,13,14,15]. Wu et al.[3] reported on a patient with type 2 diabetes with poor blood glucose control (224 mg/dL) who visited the hospital in an acute confused state. EEG performed on the 7th day of hospitalization revealed epileptiform discharges in the left fronto-temporal region, leading to the initial diagnosis of SR-NCSE.

To compensate for such delays, point-of-care EEG, which allows for the quick and easy attachment of recording electrodes, has recently attracted attention[15]. We recently conducted a retrospective study, where EEG specialists reviewed the EEGs of 5 cases of NCSE diagnosed using Headset^TM (Nihon Kohden, Tokyo, Japan). This is the only point-of-care EEG machine available in Japan, and only 1 case was correctly diagnosed with NCSE. In the remaining 4 cases, various artifacts due to insufficient attachment of electrodes or electromyographic contamination were misinterpreted as paroxysmal discharges[16]. In an acute care hospital setting, where EEG specialists are not always on-site, the practicality of point-of-care EEG has not yet been demonstrated.

In contrast, as a result of recent advances in the treatment for acute cerebral infarction, magnetic resonance imaging (MRI), mainly diffusion-weighted imaging (DWI), is now available 24 hours a day in most acute care hospitals[4, 5, 8,9,10,11,12,13,14]. Adding arterial spin labeling (ASL) perfusion imaging to routine MRI is useful in the pathophysiological diagnosis of focal epilepsy, including NCSE, based on capturing (peri-)ictal hyperperfusion linked to seizure activity via neurovascular coupling[4, 5, 8,9,10,11,12,13,14]. In this case, it is desirable to use 3-Tesla (T) pseudo-continuous labeling ASL (3-T pCASL), which can obtain the best images[4, 5, 8, 9]. However, even with the widely used 1.5-T MRI, we have emphasized the importance of improving the diagnostic accuracy of (peri-)ictal hyperperfusion by fusing ASL images with morphological images, such as DWI and fluid attenuation inversion recovery (FLAIR) images, to increase image resolution, and to highlight the areas of increased blood flow by creating subtraction of (peri-)ictal-interictal ASL co-registered with morphological MRI (SIACOM)[8,9,10,11,12,13,14]. Additionally, because ASL is affected by blood flow velocity, an accurate assessment of cerebral blood flow (CBF) is difficult with a single conventional post-labeling delay (PLD) of 1.5 s alone. Therefore, evaluating the hemodynamics of (peri-)ictal hyperperfusion with multiple PLDs is important[5, 10,11,12,13,14].

In this report, we describe a patient with newly diagnosed type 2 diabetes. In this case, capturing the presumed ictal hyperperfusion using 1.5-T pCASL with triple post-labeling delays (PLDs) of 1.5, 1.75, and 2.0 s[10,11,12,13,14], in the field of neurological emergency, was useful in the pathophysiological diagnosis of SR-NCSE that developed in association with hyperglycemia.

Case Report

A 67-year-old woman had been undergoing annual health checkups since her early 60s and had no history of abnormal blood glucose levels. Beginning approximately 2 months prior to her symptom onset, she had started drinking 1,000-2,000 ml of sugar-containing soft drinks per day to cope with the summer heat and stress of her self-employed job.

For 2 days, she had been aware of often being unable to express herself well. On the day of presentation, she had troube typing on her cell phone for approximately 2 h. Subsequently, she was accompanied by her family to our emergency room. Her height was 152 cm and body weight 56.5 kg. Her blood pressure was 198/99 mmHg. Her consciousness level was I-2 on the Japan Coma Scale. Her language comprehension was generally good; however, she had difficulty in finding words and took a long time to speak. As expected, she was unable to type on her cell phone. However, there was no obvious finger agnosia, left-right agnosia, agraphia, or acalculia. Blood chemistry results showed a blood glucose of 334 mg/dL and a HbA1c of 13.1%. Urinalysis revealed a urine glucose level of 3+; however, ketone bodies were absent.

MRI was performed 40 min after admission. ASL at a PLD of 1.5 s showed prominent hyperperfusion in the left parietal lobe, which extended to the occipital lobe (Fig. 1A). This area of increased ASL signals narrowed at a PLD of 1.75 s (Fig. 1B) and narrowed further at a PLD of 2.0 s, although remaining as a strong high-signal area (Fig. 1C, white arrows). DWI revealed a faint high-signal area almost corresponding to the cortex of the area with increased ASL signals that remained at a PLD of 2.0 s (Fig. 1D). Furthermore, FLAIR also showed a faint high-signal area almost corresponding to the cortex in the center of the area with increased ASL signals (Fig. 1E, red arrows). At the end of the MRI, her blood pressure was 133/75 mm Hg.

Fig. 1.

(A-E) Fusion of ASL with FLAIR at PLDs of 1.5 s (A), 1.75 s (B), and 2.0 s (C), DWI (D), and FLAIR (E), performed at 40 min after admission. (F-H) SIACOM with FLAIR at PLDs of 1.5 s (F), 1.75 s (G), and 2.0 s (H). Please refer to the text for details. ASL, arterial spin labeling; FLAIR, fluid-attenuated inversion recovery; PLD, post-labeling delay; DWI, diffusion-weighted imaging; SIACOM, subtraction of ictal-interictal ASL co-registered with conventional magnetic resonance imaging.

At this point, SR-NCSE associated with hyperglycemia was highly suspected. Blood glucose control with insulin was initiated. Additionally, 500 mg of levetiracetam (LEV) was administered intravenously, followed by 1,000 mg of LEV orally. Symptoms improved within 1-2 h of starting these treatments.

EEG was performed 1 day and 20 h after admission, and no paroxysmal discharge was recorded. The amplitude of photic driving was lower in the left occipital region than in the right (Fig. 2A, red arrows).

Fig. 2.

(A) EEG with averaged reference performed at 1 day and 20 h after admission. (B-D) Fusion of ASL with FLAIR at PLDs of 1.5 s, 1.75 s, and 2.0 s (B), DWI (C) and FLAIR (D), taken at 3 months after admission. ASL images of (B) show the same level as the middle row in Fig. 1A-C. Please refer to the text for details. EEG, electroencephalography; ASL, arterial spin labeling; FLAIR, fluid-attenuated inversion recovery; PLD. post-labeling delay; DWI, diffusion-weighted imaging

As blood glucose level was controlled well with short-term insulin therapy and dietary management, insulin administration was discontinued. Fasting blood glucose was maintained at 98-119 mg/dL and HbA1c at 6.5-6.6% with oral administration of 5 mg linagliptin. She was diagnosed with acute exacerbation of type 2 diabetes mellitus due to excess consumption of soft drinks and stress.

A follow-up EEG performed 3 months later failed to reveal paroxysmal discharges, and the laterality of photic driving improved. MRI showed resolution of the ASL hyperperfusion (Fig. 2B). Similarly, the high-intensity areas on DWI and FLAIR also disappeared (Fig. 2C, D).

SIACOM was created based on this interictal ASL image as previously described[6, 10,11,12,13,14]. The area of increased CBF was slightly narrower than that of the presumed ictal ASL (Fig. 1F-H). However, similar to presumed ictal ASL, the area became narrower with slower PLDs. Moreover, the area where the CBF remained at a PLD of 2.0 s (Fig. 1H, white arrows) almost coincided with the area depicted as having a high-signal intensity on FLAIR during the presumed ictal period (Fig. 1E, red arrows).

Antiseizure medication (ASM) was continued for 1 year, and then tapered over 2 months while performing repeated EEGs. Finally, it was discontinued. One year and 7 months after onset, NCSE have not recurred.

Discussion

In many acute care hospitals including ours, only 25.6% of the time is allocated for consultation, with 74.4% outside of consultation hours. Considering public holidays, approximately 3/4 of the time is non-consultative. During these hours, EEG cannot be performed immediately[8,9,10,11,12,13,14]. Our patient visited the hospital on a Saturday afternoon, and EEG could not be performed until approximately 2 days later after the start of the week.

In contrast, MRI was performed approximately 40 min after arrival. However, when conventional morphological MRI alone was used, the only positive findings were high intensity on DWI and FLAIR in the left parietal lobe, making it difficult to differentiate it from acute or subacute cerebral infarction, or encephalitis. By adding ASL with triple PLDs, it enabled us to capture presumed ictal hyperperfusion, which led to the initial suspicion of NCSE[4, 5, 8,9,10,11,12,13,14]. Further, if we consider the high-signal area on FLAIR and DWI as gliotic lesion, epileptogenic lesion, and T2 shine-through effects, respectively, it also makes it possible to differentiate it from structural focal NCSE, since the blood flow in the lesion itself has increased[4, 5, 10,11,12,13,14].

On the 1.5-T pCASL with triple PLDs of 1.5, 1.75, and 2.0 s, the high-signal area that remained until PLD of 2.0 s was diagnosed as the area with the highest increase in CBF[12,13,14,15]. In this area, the prolonged increase in metabolism due to NCSE led to relative ischemia, indicating that the blood supply could not meet the demand of increased metabolism. Additionally, excessive release of excitatory amino acids and influx of Ca ions into cells would have occurred, resulting in cytotoxic edema, which was diagnosed as showing high signals on DWI and even on FLAIR[4, 5, 12, 13]. Another characteristic is that these high-signal areas are limited to the cortex where the neurons are present[4, 5]. In other words, this high-signal intensity was not a structural abnormality that caused NCSE, but a temporary result of NCSE. Furthermore, because the left parietal lesion on DWI was not a vasogenic edema but a cytotoxic edema, posterior reversible encephalopathy syndrome could be ruled out[9]. In our patient’s MRI, this high signal intensity disappeared 3 months after hyperglycemia control and NCSE improvement. The above-mentioned course is consistent with SR-NCSE due to hyperglycemia.

Regarding the distribution of the (peri-)ictal hyperperfusion in SR-NCSE, Thomas et al.[2] showed increased blood flow in one or both frontal lobes using single-photon emission computed tomography with ^99mTc hexamethylpropylene amine oxime. We also reported increased blood flow in one or both frontal lobes using 3-T pCASL[4, 5] and 1.5-T pulsed ASL[6]. In the present case, increased blood flow was observed in the left parietal lobe. Regarding the hemodynamics of the (peri-)ictal hyperperfusion in SR-NCSE, 3-T pCASL with PLDs of 1.5 and 2.5 s showed a gradual flow type with a slow increase from a PLD of 1.5 s to a PLD of 2.5 s[5], whereas the present case showed a fast flow type with maximum signal intensity at a PLD of 1.5 s, which was attenuated at slower PLDs of 1.75 and 2.0 s on 1.5-T pCASL. Just as the causes and pathophysiology of SR-NCSE are heterogeneous[2], we believe that the distribution and hemodynamics of the (peri-)ictal hyperperfusion of SR-NCSE are also heterogeneous.

There is no consensus on how long ASM should be continued for the cases with SR-NCSE, as there is no consensus for acute symptomatic seizures including acute symptomatic status epilepticus, which are used almost synonymously, even though they often occur in association with acute lesions[14]. In this case, it would have been possible to discontinue ASM once blood glucose control had improved. However, the possibility of non-lesional elderly onset epilepsy[11] could not be completely ruled out. As a result, we discussed with the patient and decided to continue ASM for 1 year, which was then gradually tapered, and finally discontinued.

To date, few reports have been published on the diagnosis of SR-NCSE using 1.5-T ASL[6]. Additional cases should be investigated in the future. However, similar to structural focal NCSE, MRI including ASL has the advantage of being performed a more timely manner than EEG during neurological emergencies. Our experience supports the wider use of ASL in the future, and we believe that ASL is useful in diagnosing SR-NCSE, when EEG cannot be performed.

Acknowledgments

We thank Ryoji Shiraki and colleagues at the Hachisuga Hospital, for supporting our study. We thank Editage for editing the manuscript.

Informed consent

The authors confirm that written informed consent was obtained from the patient. Ethical approval was obtained from the Institutional Review Board of the Hachisuga Hospital (No. 22-1).

Conflict of Interest Disclosure

The authors declare that they have no conflicts of interest.

References

Corresponding author

Register with J-STAGE for free!