Vasospasm secondary to responsive neurostimulator placement: a previously unreported complication. Illustrative case

Abstract

BACKGROUND

The Responsive Neurostimulation (RNS) system is an implantable device for patients with drug-resistant epilepsy who are not candidates for resection of a seizure focus. As a relatively new therapeutic, the full spectrum of adverse effects has yet to be determined. A literature review revealed no previous reports of cerebral vasospasm following RNS implantation.

OBSERVATIONS

A 35-year-old man developed severe angiographic and clinical vasospasm following bilateral mesial temporal lobe RNS implantation. He initially presented with concerns for status epilepticus 8 days after implantation. On hospital day 3, a decline in his clinical examination prompted imaging studies that revealed a left middle cerebral artery (MCA) stroke with angiographic evidence of severe vasospasm of the left internal carotid artery (ICA), MCA, anterior cerebral artery (ACA), and right ICA and ACA. Despite improvements in angiographic vasospasm after appropriate treatment, a thrombus developed in the posterior M2 branch, requiring mechanical thrombectomy. Ultimately, the patient was stabilized and discharged to a rehabilitation facility with residual cognitive and motor deficits.

LESSONS

Cerebral vasospasm as a cause of ischemic stroke after uneventful RNS implantation is exceedingly rare, yet demands particular attention given the potential for severe consequences and the growing number of patients receiving RNS devices.

The Responsive Neurostimulation (RNS) system (NeuroPace, Inc.) is a device for patients with drug-resistant epilepsy who are not candidates for resection of a seizure focus.¹ The system utilizes closed-loop seizure detection and electrical stimulation to abort detected epileptic activity.^2–5 One or two depth leads, each with four-contact electrodes, are implanted as close as possible to the epileptic focus to optimize detection and stimulation.^2–5 Over half of the patients receiving an RNS implant have at least a 50% reduction in seizures by 2 years, and approximately two-thirds have a significant lasting reduction in their seizure burden.¹

Much of the understanding behind the effectiveness of RNS seizure control comes from experience with deep brain stimulation (DBS). Although the mechanisms behind RNS are yet to be completely elucidated, it has been hypothesized that RNS discharges disrupt seizure activity through short- and long-acting mechanisms.⁶ Short-acting effects may include changes in neuron excitation, either by local excitation or inhibition or by changes in blood flow and neurotransmitter release.¹ The long-acting effect of neurostimulation, as evidenced by an increased reduction in seizure activity with increased implant duration, may involve modulation of gene expression, neuronal circuit reorganization, plasticity, and neurogenesis.^1,7–10

RNS placement is regarded as a safe and effective therapeutic. Initial reports of adverse events have focused on device-related and surgical events. The most frequent device-related adverse events have been premature battery depletion (4.3%), medical device removal (3.5%), and device lead damage (3.7%).¹ The most frequent surgery-related adverse events have been infection (5.9%) and intracranial hemorrhage (4.7%).¹ Furthermore, studies have shown that RNS implantation or stimulation can produce photopsia (4.7%–14.4%), muscle twitch (3.1%), and dizziness (2.6%).^2–4 As with other brain–machine interface systems, RNS systems are subject to complications related to human–implant interaction. We present a case of vasospasm causing a large middle cerebral artery (MCA) territory stroke after RNS implantation in the setting of no known obvious etiology.

Illustrative Case

A 35-year-old male presented with a past medical history of intractable seizures refractory to medical management and a left temporal neocortical resection in 2018. He elected to undergo RNS system (Neuropace, Inc.) placement, requiring bilateral implantation of NeuroPace leads in the mesial temporal lobes and NeuroPace generator implantation (Fig. 1). Prior to his resection, he underwent depth electrode monitoring with stereo-electroencephalography (sEEG), which identified three separate epileptogenic foci: left temporal pole, left mesial temporal lobe, and right mesial temporal lobe. On the day of his procedure, a stereotactic Leksell frame was placed on the patient’s head while he was under general anesthesia. For stereotactic targeting, a computed tomography (CT) scan was obtained and co-registered with preoperative magnetic resonance imaging. RNS leads were placed in targets using posterior, longitudinal trajectories instead of the orthogonal trajectories taken during sEEG monitoring. There was no deviation from the planned coordinates and trajectories. As described in the Supplemental Methods, this was confirmed by postoperative surgical reconstruction utilizing LeadDBS (LeadDBS V2.5; Fig. 2). The procedure was performed without complications, and postoperative imaging was stable with thin tentorial subdural blood products and trace right posterior Sylvian fissure subarachnoid hemorrhage (SAH; Fig. 3). The patient was promptly restarted on his home antiepileptic regimen; however, the postoperative course was complicated by seizures on postoperative day 1. He was placed on continuous electroencephalography (cEEG), which recorded two additional seizures. cEEG was discontinued 2 days later, and the patient was discharged on postoperative day 3.

FIG. 1.

Postoperative anteroposterior (A) and lateral (B) radiographs showing lead placement.

FIG. 2.

LeadDBS surgical reconstruction showing both leads accurately placed within the hippocampus (purple).

FIG. 3.

Postoperative axial (A), coronal (B), and sagittal (C) head CT scans showing lead placement with expected postsurgical changes. The arrows (A) show locations of SAH within the occipital lobe and frontal lobe and intraventricular blood in the right atrium. Image showing prior anterior temporal lobectomy (circle, C) and subdural blood along the tentorium (arrow, C).

On postoperative day 5, the patient presented to an outside hospital after being found unresponsive with a leftward gaze preference, intermittent roving eye movements, and jerking extremity movements. The patient continually seized during Emergency Medical Services transport to the hospital, ultimately requiring intubation. The patient was febrile, and a lumbar puncture was performed with an opening pressure of 50 cm H₂O. Cerebral spinal fluid (CSF) studies showed the following: protein, 132 mg/dL; red blood cells (RBCs), 118,000/cumm; white blood cells (WBCs), 4183/cumm, and glucose, 35 mg/dL. He subsequently received broad-spectrum intravenous (IV) antibiotics. A head CT completed at the outside hospital was stable compared with postoperative imaging.

The patient was transferred to our hospital for further management and workup of status epilepticus. Upon his arrival, cEEG was initiated and his home antiepileptics were continued. Upon neurological examination, his Glasgow Coma Scale score was 7T with nonlateralizing purposeful movements in his extremities. A repeat head CT was unremarkable, with the previous small SAH remaining stable compared with that seen in Fig. 3. The working diagnosis at the time was meningitis, although it was unclear if the CSF profile was reactive from recent surgery. The patient was managed with aggressive fever management and IV antibiotics while we awaited bacterial culture results. When the patient’s fever did not improve, he was started on IV amphotericin, and a repeat lumbar puncture was performed, demonstrating the following: opening pressure, 46 cm H₂O; protein, 96; RBCs, 214; WBCs, 241; and glucose 49; and negative CSF cultures and Gram stain. Following our negative cultures and those from the outside hospital, an infectious etiology was considered less likely.

On hospital day 3, the patient developed a new right hemiparesis prompting repeat head imaging. Head CT angiography showed widespread vasospasm. We obtained a transradial catheter angiogram (TRCA), which showed severe vasospasm in the left internal cerebral artery (ICA), MCA, anterior cerebral artery (ACA), and right ICA and ACA (Fig. 4A). The patient was treated with an injection of 20 mg verapamil into the left ICA, 10 mg into the right ICA, and 10 mg into the left vertebral artery. The following day, transcranial Doppler (TCD) ultrasonography showed markedly increased blood-flow velocities (left Lindegaard ratio of 5.8) in the bilateral MCA and ACA territories, prompting a repeat TRCA, which revealed a left posterior M2 branch occlusion (Fig. 4C). Blood flow was successfully restored by direct aspiration technique. During the procedure, severe underlying vasospasm was discovered and treated with balloon angioplasty and intra-arterial verapamil (Fig. 4D). After the procedure, CT imaging revealed an evolving, large hypodense area in the circulation of the left MCA (Fig. 5B). On hospital day 7, TCD ultrasonography showed multiple elevations in blood-flow velocity. On repeat TRCA, he was found to have mild vasospasm in the anterior circulation and moderate bilateral vasospasm in the posterior circulation, which was treated with verapamil. The patient’s clinical examination improved after the vasospastic episodes subsided, and he was extubated on hospital day 11 and discharged to a rehabilitation center on day 19. At discharge, the patient was hemiparetic on the right and was oriented to person and place. The patient was discharged to an acute rehabilitation center with a modified Rankin scale score of 4. The patient has been unable to follow up with our clinic and is under the care of an outside neurologist.

FIG. 4.

Left supraclinoid ICA showing severe vasospasm (A, arrow), resolution of vasospasm after the administration of verapamil (B, arrow), occlusion of the posterior M2 branch (C, arrow), and reconstitution of the posterior M2 branch on post-thrombectomy angiography (D, arrow).

FIG. 5.

A: CT perfusion scan showing reduced cerebral blood flow (areas of blue/black) in the left MCA distribution. B: Head CT on hospital day 6 showing a large hypodensity in the distribution of the left MCA (arrow).

Discussion

Observations

The most common complication after NeuroPace implantation is infection (5.9%) and rarely lead explantation or implant-related hemorrhage.¹ The rate of these complications is similar to that of DBS devices.¹¹ Symptomatic DBS-related hemorrhage has been reported to occur in 1.2% of patients.¹² Risk factors for hemorrhage in DBS placement include a history of hypertension, older age, sulcus or ventricle penetration during lead placement, and implantation without microelectrode recording guidance.¹³

Further review of the literature did not reveal any cases of vasospasm after NeuroPace placement or status epilepticus as a trigger for vasospasm, even though there were several cases of hemorrhage postoperatively. Likewise, stroke is exceptionally rare in the DBS literature, with an ischemic stroke occurring in approximately 2% of patients with globus pallidus internus implants and 0.2% in subthalamic nucleus implants.¹⁴ The etiology of stroke in these cases is hypothesized to be electrical stimulation–induced vasospasm, mechanical disruption of blood vessels causing rupture or vasospasm, and mechanical compression due to lead placement or edema.¹⁵ Although the association between vasospasm in aneurysm rupture has been thoroughly investigated, other causes are less understood. Electrically induced vasospasm has also been studied in animal models as a mechanism for hemorrhage control. A rat model showed that the femoral artery constricts by 75% when stimulated using 150 V with 100-millisecond pulses at 10 Hz for 30 seconds.¹⁶ However, this study used a far higher voltage than DBS or NeuroPace devices. Another experiment demonstrated 41% arterial constriction while using far lower voltage settings of 20 V at 10 Hz for 5 minutes and further suggested a neural mechanism as the cause of vasoconstriction.¹⁷

In our case, the presence of severe, generalized vasospasm is unusual given the minimal presence of postoperative subarachnoid blood and the fact that the device had not been activated. The utilization of a different trajectory than the sEEG electrodes lessened the risk of vascular injury. Although several angiograms remained negative, a pseudoaneurysm from occult vascular injury, which bled briefly and subsequently coagulated to become imaging occult, is a possible cause of the SAH. Importantly, the patient’s vasospasm occurred within the range of peak RBC breakdown and inflammatory cell activity, approximately 7 days post-SAH, coinciding with peak CSF hemoglobin in animal studies.^18,19 It has been postulated that neuroinflammation after SAH may play an important role in delayed cerebral ischemia.²⁰ As inflammatory cells migrate into the subarachnoid space to digest RBCs, they produce several vasoactive and inflammatory mediators. Although this process is necessary for aiding neuronal recovery, the byproducts of RBC breakdown, specifically the production of oxyhemoglobin, bilirubin oxidation products, and inflammatory cytokines, have been linked to vasospasm.^20–22 The final cultures from all CSF analyses were negative for infection. Thus, the patient’s CSF findings of an elevated WBC count and negative CSF cultures were indicative of an inflammatory process. Moreover, the patient’s elevated lumbar puncture opening pressure indicates that increased intracranial pressure (ICP) may have contributed to the severity of the patient’s vasospasm. It has been shown that an increased ICP is not only related to vasospasm but is also predictive of a poor prognosis.²³ Some authors have suggested that increased ICP in the setting of vasospasm is caused by ischemia-related cerebral edema.²⁴ Moreover, an animal study by Farrar²⁵ demonstrated a positive correlation between ICP and vasospasm, such that increased ICP worsens chronic vasospasm. Another possibility is an immune reaction to the leads, which mediates the release of inflammatory and vasoactive mediators.

Lessons

Although stroke and vasospasm are exceptionally rare complications of RNS or DBS, clinicians should remain aware of the possibility of this complication given its devastating sequelae. Patients with a history of previous vasoconstrictive episodes should be monitored closely. To expedite treatment, patients with surgically induced subarachnoid blood should be followed closely and educated on the potential signs of vasospasm.

Here, we presented a case of vasospasm after NeuroPace implantation. Vasospasm following aneurysmal SAH is a well-established complication, but vasospasm can occur regardless of the cause of hemorrhage. Although the exact etiology of vasospasm is unclear in this case, physicians should be aware of the possibility that electrical stimulation, mechanical irritation, or subarachnoid blood could trigger vasospasm.

Disclosures

Dr. Lena reported personal fees from Stryker outside the submitted work.

Author Contributions

Conception and design: Vandergrift, Miller, Saway, Rowland, Lena. Acquisition of data: Rogowski, Miller, Saway, Rowland, Lena. Analysis and interpretation of data: Miller, Saway, Rowland, Lena. Drafting of the article: Rogowski, Miller, Saway, Rowland. Critically revising the article: all authors. Reviewed submitted version of the manuscript: Rogowski, Miller, Saway, Wessell, Rowland, Lena. Approved the final version of the manuscript on behalf of all authors: Vandergrift. Statistical analysis: Rowland. Administrative/technical/material support: Saway, Rowland. Study supervision: Saway, Rowland, Lena.

Supplemental Information

Online-Only Content

Supplemental material is available with the online version of the article.

Supplemental Methods. https://thejns.org/doi/suppl/10.3171/CASE22435.