Treat the Patient,Not the Test When a Hospitalized Patient in Status Epilepticus Transitions to Comfort-Focused Goals of Care

Abstract

Status epilepticus is a common and under-recognized cause of unconsciousness among hospitalized patients. It can clinically mimic delirium and other causes of acute mental status change, especially when clinically relevant seizure activity is not appreciated on physical examination. While the successful treatment of status epilepticus may require anesthetic dosing of antiepileptics such as barbiturates, these seemingly drastic therapeutic measures are justified when goals of care are life prolonging as they may allow a patient to regain consciousness and meaningfully interact with loved ones. However, medical burden from electroencephalogram (EEG) monitoring and other care associated with its diagnosis and treatment can contribute to distress for patients who may be dying from a comorbid illness. Furthermore, when goals of care transition to comfort, care challenges can result regarding the ongoing management of barbiturates or other sedatives, previously prescribed to treat the status epilepticus. In this case study, the lack of clinically significant seizure activity led us to conclude that the discontinuation of a barbiturate infusion and the EEG monitoring was the clinically appropriate approach for a dying patient with comfort goals of care and nonconvulsive status epilepticus.

Background

Status epilepticus is an under-recognized cause of unconsciousness and mortality among critically ill patients.^1–3 Patients with a comorbid life-limiting illness may be at heightened risk for status epilepticus.¹ The clinical presentation can be subtle (e.g., nystagmoid eye movements or minimal facial twitching) or even unnoticeable.⁴ When continuous or recurrent seizures occur without convulsive activity, they may be identifiable only with electroencephalogram (EEG) monitoring, a condition known as nonconvulsive status epilepticus (NCSE).⁴ Although status epilepticus is associated with high morbidity and mortality rates,^2,5 NCSE may have better outcomes than convulsive status epilepticus if properly recognized and treated.^4,6 There are case reports of patients admitted to a palliative care unit who regained the ability to communicate meaningfully with loved ones and function independently after successful management of NCSE.^7,8 Clinical vigilance and EEG monitoring are recommended to better identify and treat NCSE.^9,10

However, increased screening for NCSE amplifies the need for more defined care strategies. Goals of care for patients with NCSE who are terminally ill from a separate illness may be challenging to define and manage. There has been little written in the literature about status epilepticus management strategies for patients who transition to palliative-based care goals, even though the care challenges are complex and unique. For example, anesthetic dosing levels of benzodiazepines, barbiturates, or propofol (pharmacotherapies which palliative care clinicians may more commonly associate with the intent of delivering palliative sedation,^11,12 rather than life prolongation) may be utilized as life-prolonging therapies for status epilepticus. These therapies are associated with their own morbidity and mortality risks. If a decision to transition to comfort-focused goals of care occurs, the continued administration of these medications may not be clinically or ethically justified especially if seizures are appreciated only through EEG monitoring.

Another case report has discussed the challenges of managing status epilepticus in a hospice inpatient setting,¹³ but in this article, we will review and discuss our clinical reasoning underlying the management of a hospitalized patient with NCSE whose goals of care transitioned to comfort. Through this analysis, we aim to illustrate the importance of defining the intention of status epilepticus treatment and the goals of care with surrogate decision makers early in the patient's presentation. We will also highlight the competing care challenges when managing status epilepticus in a patient with a comorbid terminal illness.

Case

A 64-year-old woman with congestive heart failure, and the sequelae of cognitive impairment and epilepsy, refractory to a multidrug antiepileptic regimen, from a prior anoxic brain injury was admitted to a neurologic intensive care unit (NICU) with a Glasgow Coma Scale of 3 after being found unresponsive in her home. She was intubated and placed on mechanical ventilation for airway protection. No significant metabolic, medication, intracranial, or toxin-related cause for her altered mentation was identified. Continuous EEG monitoring revealed diffuse slowing and near continuous bisynchronous bilateral-anterior high amplitude epileptiform sharp waves at a frequency of 1–2 Hz. These epileptiform discharges demonstrated a waxing and waning frequency consistent with an ictal (i.e., seizure) pattern.

Despite an extended trial of artificial nutrition through a nasogastric (NG) feeding tube, ventilatory support through an endotracheal tube, and a multiagent antiepileptic regimen (Table 1) which included an intravenous (IV) pentobarbital infusion, her continuous EEG monitoring continued to demonstrate severe encephalopathy and NCSE. Besides spontaneous chewing of the endotracheal tube and occasional eyelid twitching, no visible signs of clinical seizure activity or physical discomfort were noted.

Table 1.

Inpatient Antiepileptic Regimen

Antiepileptic agent	Maximal daily dose	IV or NG route
Pentobarbital	2100 mg	IV
Phenobarbital	120 mg	IV
Phenytoin	600 mg	IV
Lacosamide	400 mg	IV
Divalproex	3000 mg	IV
Gabapentin	1800	NG
Lorazepam	6 mg	NG

IV, intravenous; NG, nasogastric.

Several weeks into the hospitalization, a goals of care discussion ensued with the NICU clinicians and the patient's children who were making medical decisions by family consensus. The NICU clinicians conveyed that placement of a tracheostomy and a gastric feeding tube would be needed for life prolongation, as well as indefinite placement in a nursing home or long-term acute care facility, since prompt neurologic recovery to an independent functional status was unlikely. Her children presented clear evidence that in these circumstances the patient would not want to pursue these measures. They asked the NICU team to withdraw artificial nutrition and ventilatory support in the hopes that she could be transferred out of the NICU to a more peaceful environment for her remaining days.

The inpatient palliative care team was consulted for guidance on end-of-life management. At the time of the initial consultation, the NICU team expressed conflicting concerns about the IV pentobarbital infusion which had been infusing for approximately two weeks. On the one hand, concern was expressed that reducing or discontinuing the pentobarbital infusion could result in clinical seizures given the electrographic evidence for seizure activity. On the other hand, there was concern that continuing a pentobarbital infusion while removing ventilatory support could artificially hasten death.

Case resolution

No clear evidence of physical or psychological distress was identified nor postulated to be the result of her NCSE. Instead, the primary source of distress appeared to be iatrogenic through the erosion of the patient's perceived dignity from the medical interventions required to appropriately monitor and manage her NCSE. The family prioritized the minimization of medical interventions over the maximization of eventual NCSE control and, therefore, asked if the headpiece required for EEG monitoring be removed. Neurologic checks, artificial nutrition, continuous EEG monitoring, scheduled phenobarbital and lorazepam, and her pentobarbital infusion were discontinued. Even though lacosamide, phenytoin, and divalproex were not anticipated to control her NCSE, we continued these antiepileptics as part of her comfort care plan seeing as we did not anticipate them to cause high levels of iatrogenic sedation nor was the scheduled IV administration anticipated to require additional inpatient monitoring or burden.

In light of the variable elimination half-life of pentobarbital in a critically ill patient,¹⁴ extubation and removal of the patient's endotracheal tube were delayed until there was confirmation by serum laboratory levels that the pentobarbital was no longer at therapeutic doses. Approximately 24 hours later, subtherapeutic pentobarbital levels were confirmed. Therefore, clinicians were confident that the patient's obtunded mental status was not the result of a medically induced coma. A comfort care plan was agreed upon between clinicians and family. This included removal of her NG tube and discontinuation of gabapentin, as well as scheduled IV antiepileptics, anticipated to incur a high degree of artificial sedation (lorazepam and phenobarbital). IV boluses of 2 mg of lorazepam ± reinstitution of her pentobarbital IV infusion would be administered only if clinical signs of seizure activity were observed. Bolus one-time doses of 10 mg of IV diazepam and 10 mg of IV morphine were administered about 30 minutes before extubation to reduce the chances of respiratory distress and/or clinically relevant seizure activity. After the extubation, she was transferred to a general medicine bed. Her family remained at her bedside until her death 1 day later. Slight spontaneous nonrhythmic movements were noted on occasion, but otherwise no signs of distress were noted. In her final 24 hours of life, morphine for observed respiratory distress, acetaminophen for fever, and glycopyrrolate for oral secretions were the only administered medications.

Discussion

In this case, the patient exhibited multiple predictors of a poor functional outcome, including comorbid medical illnesses, a prolonged duration of a drug-induced coma, a prolonged need for mechanical ventilation, and a prolonged period of severe encephalopathy, as seen on continuous EEG.¹⁵ Even though the successful treatment of NSCE in isolation could potentially allow a patient to regain consciousness and meaningfully interact with family members before death, for this patient, her known sequelae from a prior anoxic brain injury made this type of recovery unlikely. Furthermore, the family offered an informed and clear expression of the patient's medical wishes knowing the burdens and prolonged immobility required from life-prolonging treatments. This clarified that the primary goal of care should not be restoration of cognitive and physical function, but rather preservation of dignity, minimization of suffering, and restoration of a less medicalized care environment. Clear consensus was achieved that a transition to comfort care was consistent with a humane and ethical approach described by other authors in the published medical literature.¹³

There was less consensus about whether the withdrawal of the pentobarbital infusion would lead to increased psychological or physical distress through increased epileptic activity. In our clinical assessments, her epileptic activity was more consistent with NCSE (i.e., there were no overt motor manifestations of her seizures). As far as we could ascertain, there was no clearly identified pathophysiologic mechanism for physical, psychological, spiritual, or existential suffering to occur from a stuporous mental state induced by NCSE. Because of the family's wishes, the principle therapeutic aim shifted away from preventing all seizure activity (as observed through continuous EEG monitoring) to preventing or addressing overt clinical seizures through the patient's dying process. Considering the family's clearly conveyed sense of the patient's substituted care preferences, we hypothesized that distress could result from the continued use of EEG, nursing, and medical monitoring by contributing to a more medicalized environment. Hence, our comfort care plan prioritized the minimization of medical interventions which could lead to discomfort, especially ones which would be disruptive or esthetically unappealing to the care environment. Through this clinical reasoning, we removed the electrodes required for EEG monitoring and discontinued routine neurologic checks and mechanical ventilatory support and removed the endotracheal and NG tubes with the goal of transferring her out of the NICU care setting.

Through multidisciplinary discussions, a consensus was reached to discontinue the pentobarbital infusion. Because we could not clearly identify any clear suffering from her NCSE, the continued use of a pentobarbital infusion at doses meant to medically induce a coma was felt to be disproportionate to the symptoms we were aiming to alleviate. Considering that her prognosis was estimated to be only hours to days after extubation, we felt it was likely that seizures would remain subclinical until her death. The timing of extubation relative to discontinuation of pentobarbital remains an area for further discussion. Given the goals of care and her anticipated prognosis after extubation, one can argue that waiting to verify a subtherapeutic level before extubation was not necessary. Perhaps a bioethics consultation would have been helpful in resolving this clinical question.

Another potential critique in our management was the selection of diazepam as the prophylactic benzodiazepine before extubation. One author suggested lorazepam, midazolam, or clonazepam as preferred alternatives for NCSE, considering that the lipophilic character of diazepam could compromise its antiepileptic effect through rapid redistribution in body fat.¹

The patient did not demonstrate any visible signs of convulsive activity through her imminent dying process. There was also a sense shared among the treating clinicians and the patient's family that the discontinuation of the continuous EEG monitoring and the removal of the headset offered a certain liberation from its influence on medical decisions and family interactions. Although we may never know the patient's experienced symptomatology after extubation, our best indirect corollary was grateful feedback from her bereaved family that she appeared like their mother again for a short period of time before her death.

Conclusion

This case exposed many unanswered clinical questions regarding the management of NCSE in a patient whose goals of care transitioned to comfort care which are worthy of further investigation. Is NCSE commonly mistaken as terminal delirium? Do patients physically, psychologically, or spiritually suffer from NCSE? What is the best therapeutic approach for patients with refractory seizures at the end of life which minimizes the burden of medical intervention while maximizing the chances for meaningful interactions with loved ones? Despite these unresolved questions, we identified the following medical and ethical take-home points from this case analysis:

While continuous EEG monitoring is a vital tool for status epilepticus, ongoing EEG monitoring in dying patients who transition to comfort goals of care may engender medical interventions that are not symptom based.

De-escalation of anesthetic agents may be necessary care component when NCSE patients transition to comfort care, as their continued use raises ethical concerns regarding proportionality.

A multidisciplinary approach involving ongoing input from palliative care, critical care, and neurology is recommended for dying patients with status epilepticus.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Towne

, Waterhouse

, Boggs

, et al.: Prevalence of nonconvulsive status epilepticus in comatose patients. Neurology, 2000; 54:340–345.

Datar

: New developments in refractory status epilepticus. Neurol Clin, 2017; 35:751–760.

Dham

, Hunter

, Rincon

: The epidemiology of status epilepticus in the United States. Neurocrit Care, 2014; 20:476–483.

Power

, Gramstad

, Gilhus

, Englesen

: Adult nonconvulsive status epilepticus in a clinical setting: Semiology, aetiology, treatment, and outcome. Seizure, 2015; 24:102–106.

Legriel

, Azoulay

, Resche-Rigon

, et al.: Functional outcome after convulsive status epilepticus. Crit Care Med, 2010; 38:2295–2303.

Standley

, Abdulmassih

, Benbadis

: Good outcome is possible after months of refractory convulsive status epilepticus: Lesson learned. Epilepsia, 2012; 53:e17–e20.

Samala

, Paral-Metz

, Davis

: Nonconvulsive status epilepticus in a palliative care unit: When delirium is a seizure. Am J Hosp Palliat Med, 2015; 32:243–247.

Lorenzl

, Mayer

, Feddersen

, et al.: Nonconvulsive status epilepticus in palliative care patients. J Pain Symptom Manage, 2012; 40:460–466.

Maganti

, Gerber

, Drees

, Chung

: Nonconvulsive status epilepticus. Epilepsy Behav, 2008; 12:572–586.

10.

Marcuse

, Lancman

, Demopoulos

, Fields

: Nonconvulsive status epilepticus in patients with brain tumors. Seizure, 2014; 23:54–57.

11.

Maltoni

, Scarpi

: Palliative sedation in end-of-life care and survival: A systematic review. J Clinical Oncol, 2012; 30:1378–1383.

12.

Maiser

, Estrada-Stephen

, Sahr

, et al.: A survey of hospice and palliative care clinicians' experiences and attitudes regarding the use of palliative sedation. J Palliat Med, 2017; 20:915–921.

13.

Droney

, Hall

: Status epilepticus in a hospice inpatient setting. J Pain Symptom Manage, 2008; 36:97–105.

14.

Ehrnebo

: Pharmacokinetics and distribution properties of pentobarbital in humans following oral and intravenous administration. J Pharm Sci, 1974; 63:1114–1118.

15.

Hocker

, Britton

, Mandrekar

, et al.: Predictors of outcome in refractory status epilepticus. JAMA Neurol, 2013; 70:72–77.