Abstract
Keywords
Nonconvulsive generalised status epilepticus (NGS) occurring following electroconvulsive therapy (ECT) is a rare event. Four cases have appeared in the English literature since it was first reported in 1980 [1–4]. Here we present a patient who developed NGS during a course of ECT, having been treated with haloperidol, paroxetine, and clonazepam.
Case report
The patient, Y, was a single 40-year-old woman who had been admitted medically to a major city hospital following an overdose of nortriptyline. She had been treated unsuccessfully over the previous 12 months for a major depressive episode. Because of concerns about her worsening mood and suicidality she was transferred, as a voluntary patient, to the inpatient psychiatric unit. On transfer her medications were haloperidol 1.5 mg nocte, nefazodone 50 mg bd, temazepam 20 mg prn nocte, and clonazepam as needed for agitation.
Y had been reviewed 5 years previously by the neurology service for investigation of migraines. A computerised tomography (CT) of her head at the time was normal. She had also suffered a minor head injury following a motor vehicle accident 15 years earlier. She had no personal or family history of epilepsy. She denied significant drug or alcohol use.
Following transfer nefazodone was stopped and paroxetine commenced, 20 mg daily. Over the next 2 weeks her mental state continued to worsen. A course of ECT was recommended. Informed consent was obtained. No changes were made to her medications. Over the next 10 days Y received a total of five ECT treatments. A Thymatron DGx (Somatics, Lake Bluff, IL, USA) was used. The anaesthetic agents used were methohexitone and suxamethonium. Because of incomplete muscle relaxation, seizures were visible obviating the need for a cuffed limb. Electroconvulsive therapy was administered bilaterally. The initial electrical stimulus of 25%% of full (126 mC) was according to unit protocol. It was increased to 30%% at the third treatment and 35%% (176.4 mC) at the last as seizure lengths shortened. Seizure duration as monitored by single channel electroencephalogram (EEG) varied between 32 s to 51 s while physically observed seizures ranged from 20 s to 40 s. The Seizure Energy Index as measured by the Thymatron DGx ranged between 640 and 1130. The Postictal Suppression Index fell below the recommended 74%% [5] at the third and fifth treatments, it was not recorded for the fourth. Apart from a mild headache after the first treatment there appeared to be no other significant problems. Throughout the course Y's mood improved steadily. Following the fifth treatment most of her depressive symptoms had resolved. She had required no clonazepam from the time of her first treatment.
The day after her fifth treatment, nursing staff noted Y to be confused and behaving unusually. An examination revealed bilateral up-going plantar reflexes. Investigations including full blood count, urea and electrolytes, liver function tests, chest X-ray and CT head were all normal. The following day her plantar reflexes had returned to normal.
When seen by the treating team 2 days following her last ECT Y stated she was ‘not feeling well’. During the interview she was poorly attentive. Marked motor retardation and a paucity of movement were evident. Her speech was slowed with latency of reply and poverty of content. She was disorientated to time and place, and did not recognise staff. She was unable to comment on her mood. Her affect was blunted.
An urgent neurology review and EEG were requested. The EEG report noted ‘frequent bursts of high amplitude, generalised, slow wave activity and unequivocal spikes (were) intermixed with these bursts, although for most of the recording they are relatively unobtrusive. Overall the EEG suggests that the patient is in non-convulsive status epilepticus’. She was immediately transferred to the neurology ward. Intravenous diazepam (a single 5 mg dose) led to a definite transient improvement in her mental state. Phenytoin was loaded intravenously and continued orally. Paroxetine and haloperidol were stopped. The following day she remained confused but again showed a partial improvement, both clinically and on EEG, when given intravenous diazepam.
Two days later Y was transferred back to the psychiatric unit. Her mood was euthymic and, though she had only vague recollections of the previous 5 days, she was cognitively intact. Phenytoin was changed to carbamazepine in the hope this would maintain her mood. She remained well and was discharged 3 weeks later with community follow up. A repeat EEG 6 weeks following the onset of NGS was reported as ‘mildly abnormal… because of an excess of theta and delta activity arising in a paroxysmal fashion over the temporal regions, particularly on the left’, no definite epileptiform activity was identified. At follow up 3 months later she had no significant mood or cognitive symptoms and no recurrence of seizures.
Discussion
Non-convulsive generalised status epilepticus (NGS) has been defined as a state lasting more than 1 h, characterised by a slowness in behaviour and mentation, confusion, and sometimes stupor (or coma), accompanied by generalised continuous or nearly continuous activity in the form of spike waves, polyspike waves, or more complex discharges on the EEG. Clinically, patients' mental states have varied from slowed and confused to alert and floridly psychotic [6].
A number of factors have been identified as possible precipitants of NGS [1],[6]. Concurrent paroxetine use [7], hypno-sedative withdrawal and seizure threshold-reducing medications such as neuroleptics [8] have all been implicated in prolonged seizures following ECT. Lithium [1], thioridazine [4], and haloperidol [2] were felt to be aetiologically important in the cases in the literature. In the above case, concurrent haloperidol and paroxetine use, and probable benzodiazepine withdrawal, may all be implicated although their relative importance remains unclear. Despite medications such as lithium, antipsychotics and antidepressants having been associated with adverse effects following ECT, the supporting literature is not conclusive and all have been used safely routinely [9].
Diagnosis of NGS following ECT is difficult and must be made with caution [10],[11]. Acute confusional states following ECT are not uncommon and, although usually more gradual in onset, can be difficult to differentiate from NGS [10],[11]. Causes include the frequently seen transient post-ECT confusion, post-ECT delirium [11] and delirium due to other general medical disorders. Also, EEG changes developing during a course of ECT, with marked increases in delta and theta activity lasting weeks [12], as well as the EEG changes that accompany various causes of delirium [11], may be difficult to differentiate from NGS and should be interpreted with care [10]. These problems, as well as confounding medications, have led to the diagnosis having been questioned in the reported cases [11]. The use of intravenous diazepam to separate NGS from other encephalopathies with similar clinical presentations and EEG changes, with an improvement in the level of consciousness seen in patients in NGS, has been advocated [6]. This use of an anti-convulsant agent to differentiate patients in NGS from an ECT-induced delirium is felt by some investigators to be definitive [11].
NGS presenting with symptoms of psychosis has been reported [6] and in the cases reported has been mistaken for a relapse of a psychiatric disorder. In one case, the patient's decline was attributed to a recurrence of her previous psychotic symptoms [2]. In another, a psychotic decompensation was considered in the differential diagnosis [3]. In both cases, however, the patients showed fluctuating levels of consciousness and confusion.
There is little information about the treatment of NGS occurring after ECT. In the cases reported all were given intravenous diazepam followed by an oral anticonvulsant agent, phenytoin or valproic acid [1–4]. There is no clear guide as to how long these need to be continued. In the case reported by Varma et al. [2] the patient ‘later had a recurrence of the paroxysmal altered mental state’ associated with a low valproate level, which responded to an increased valproate dose.
There is no information regarding recurrence. Although the further use of ECT following NGS has not been reported, it has been used safely after ECT-induced convulsive status under anticonvulsant cover [13]. In the patient reported by Weiner et al. [1], and our case, medications that were felt to be contributing were stopped following diagnosis. The lithium stopped in the case reported by Weiner et al. was recommenced safely weeks after the resolution of the NGS. In the case reported by Grogan et al. [3] antipsychotic medication was safely used at an unspecified time following the episode of NGS.
The incidence of cognitive deficits following NGS is also not known. Guberman et al. [6] reported a series of patients who had suffered recurrent episodes of NGS from a number of causes. In this group it was noted that ‘despite recurrent bouts of status, there was no evidence of long-term intellectual, memory, or behavioural deterioration’. The data in support of this was not documented and it remains an area of concern.
In summary we present a further case of NGS occurring following ECT. Although it is a rare complication of ECT and requires careful consideration in its diagnosis, it is easily treated and should be included in the differential diagnosis of any patient displaying unusual behaviour of sudden onset following ECT.