Identification of predictors of post-ictal delirium after electroconvulsive therapy


Norio Yasui-Furukori, MD, PhD, Department of Neuropsychiatry, Graduate School of Medicine, Hirosaki University, Hirosaki 036-8562, Japan. Email:


Aim:  Although electroconvulsive therapy (ECT) often causes post-ictal delirium (PID), to date, the specific risk factors of PID have not been described. The purpose of the present study was therefore to elucidate the predictors of PID via identification of the characteristics of patients with PID.

Methods:  ECT was conducted in 50 patients and all patients underwent more than four sessions. A sine wave or a brief-pulse square wave ECT instrument was used. After convulsions the patients' PID was monitored for 30 min. The patients were allocated into four groups based on PID severity (none, mild, moderate or severe PID). Variables, including age, gender, duration of illness, diagnosis, clinical features (psychotic or catatonic features) and stimulus waveform (sine or brief pulse square waveform), were analyzed.

Results:  Moderate to severe PID developed during the ECT sessions in 18 patients (36%). Most patients with severe delirium were successfully treated with i.v. bolus of propofol (1–2 mg/kg). Although the incidence of PID was 24% in patients without catatonic features, the incidence in patients with catatonic features was extremely high (88%; P < 0.001). Multiple regression analyses showed that the severity of PID correlated significantly with the presence of catatonic features (β = 0.428, P < 0.01).

Conclusion:  The presence of catatonic features before ECT is a predictor of PID. Propofol is useful for the treatment of PID.

THE PRIMARY GOAL during the post-ictal phase of electroconvulsive therapy (ECT) is to maintain an adequate airway until the return of spontaneous respiration and subsequent alertness. But post-ictal delirium (PID) or emergence delirium, which is an acute confusional state, can often occur during the immediate post-ictal phase in patients receiving ECT. Following seizure elicitation with ECT, some patients exhibit PID.1–5 Approximately 10% of patients develop a self-limited delirium or acute confusional state during the immediate post-ictal phase.6 The state is characterized by motor agitation, thrashing, clouded consciousness, disorientation, and poor response to commands. When PID is severe or persistent it is necessary to administer an i.v. hypnotic. To our knowledge, however, there has been no information about the specific risk factors of PID in ECT. The aim of the present study was to prospectively identify predictors of PID after ECT.



The subjects consisted of 50 inpatients (16 men, 34 women) in the Psychiatric Ward of Hirosaki University Hospital, Aomori, Japan. The mean (±SD) age was 48.0 ± 17.4 years. The mean duration of illness was 11.1 ± 11.5 years. The patients had been diagnosed according to DSM-IV-TR criteria7 and the structured clinical interview for DSM-IV axis I disorder.8 The ‘with catatonic features’ specifier was applied if the clinical picture was dominated by at least two of the following: (i) motor immobility, as evidenced by catalepsy (including waxy flexibility) or stupor; (ii) excessive motor activity (purposeless, not influenced by external stimuli); (iii) extreme negativism (motiveless resistance to all instructions or maintenance of a rigid posture against attempts to be moved) or mutism; (iv) peculiarities of movements, prominent mannerisms, or prominent grimacing; (v) echolalia or echopraxia. The ‘with psychotic features’ specifier was applied to the mental state with delusions or hallucinations regardless of the diagnosis. The patient diagnoses included schizophrenia (n = 23: paranoid, 15 patients; disorganized, three patients; catatonic, four patients; residual, one patient), mood disorder (n = 22: major depressive disorder, single episode, four patients; major depressive disorder, recurrent, 16 patients; bipolar I disorder, one patient; and bipolar II disorder, one patient) and somatoform disorder (n = 5: pain disorder, two patients; conversion disorder; two patients; hypochondriasis, one patient), with <25 points on the Global Assessment of Functioning, which is a numeric scale (0 through 100) used by mental health clinicians to rate the social, occupational and psychological functioning of adults described in the DSM-IV.7 In somatoform disorders, the target symptoms were pain and cenesthopathy.

There were 33 patients with psychotic features and nine patients with catatonic features. Details of the catatonic symptoms were as follows: (i) motor immobility or stupor, 67%; (ii) excessive motor activity, 56%; (iii) extreme negativism or mutism, 100%; (iv) peculiarities of voluntary movement evidenced by posturing, stereotypical movement, prominent mannerism, or prominent grimacing, 44%; and (v) echolalia or echopraxia, 11%. In mood disorders with catatonic features (n = 4), two patients had symptoms of (i), (ii) and (iii), and the remaining two patients had symptoms of (i) and (iii). These four patients were suffering from severe major depressive disorder.

Patients with clinically significant abnormal laboratory or electrocardiography findings, epilepsy, alcoholism or drug abuse, and clinically significant organic or neurological disease were excluded from the present study. Patients with severe physical complications were also excluded from the study. The ECT stimulus was delivered using a sine wave ECT instrument (n = 36) or a brief-pulse square wave ECT instrument (n = 14). This study was performed from February 2005 to December 2007.

Informed consent was obtained from each patient or, when necessary, the appropriate relative. The study protocol was approved by the Ethics Committee of Hirosaki University Hospital.

Anesthesia and electroconvulsive therapy

General anesthesia and paralysis were induced with propofol 1–2 mg/kg and suxamethonium 1 mg/kg, respectively. Adequate dosage of propofol was administered using a bispectral index (BIS) monitor. All patients were oxygenated before ECT with 100% oxygen by mask. The patients were monitored with a sphygmomanometer, electrocardiogram, SpO2 monitor and the BIS monitor. The electroshock stimulus was delivered by well-trained psychiatrists using a sine wave ECT apparatus or a brief-pulse square wave ECT apparatus. A sine wave ECT was conducted for 5 s at 100 V on a sine wave apparatus (C-2, Sakai Medical, Tokyo, Japan), and electrodes were placed in the traditional bilateral fronto-temporal manner. In cases in which brief-pulse square ECT was administered, patients were treated with bilateral ECT using a Thymatron System (Somatics, Lake Bluff, Ill, USA) set at a current of 0.9 A and a wave with a pulse width of 0.5 ms; electrode placement was bitemporal, and the half-age method was used to determine the energy level. For prevention of tachycardia and hypertension, diltiazem (5–10 mg) was administered before the electroshock stimulus in patients with hypertension or in elderly patients (>65 years of age).

Classification of PID severity

After convulsions the patients were monitored for 30 min and PID scores were evaluated and the level of severity was defined as follows: 0, no delirium; 1, mild delirium (it was possible to cope with the delirium with a verbal command alone); 2, moderate delirium (necessary to manually restrain the patient who exhibits excitement and thrashing) or 3, severe delirium (delirium prolonged for 2–3 min and manual restraint of the patient necessary three times, injection of a sedative required to end the delirium).

Treatment of post-ictal delirium

Patients with severe post-ictal delirium (score 3) were treated with injection of 1–2 mg/kg propofol. The dose of propofol was the same as the induction dosage. After sedation with propofol, SpO2, blood pressure and heart rate were monitored carefully. When PID was not terminated by i.v. propofol bolus, continuous i.v. infusion (4–8 mg/kg per h) of propofol was administered to the patient for 15–20 min. During propofol infusion the airway patency was maintained: spontaneous ventilation was unobstructed with supplemental oxygenation.

Assessment of the effects of ECT

In patients with schizophrenia, we evaluated the clinical response to ECT using the Brief Psychiatric Rating Scale.9 For patients with mood disorder, the Montgomery Asberg Depression Rating Scale was used.10 In patients with somatoform disorder, especially pain disorder, a visual analog scale was used. In all patients, response was defined as a reduction ≥50% from the baseline score.

Statistical analysis

Various factors, including age, gender, duration of illness, diagnosis, clinical features (psychotic or catatonic features), and stimulus waveform (sine or brief pulse square waveform), were investigated prospectively. Because a post-ictal delirium score of >2 was considered to be a clinical problem, the patients were divided into two groups: patients who had no or mild delirium (score 0–1) and those who had moderate to severe delirium (score 2–3). The clinical characteristics were compared between the two groups using t-test and Fisher's exact test. Using multiple regression analyses, risk factors associated with PID scores were determined. We regarded PID scores from 0 to 3 and clinical factors as dependent and independent variables, respectively, in multiple regression analyses. P < 0.05 was considered to be significant. All calculations were performed using SPSS version 11 (SPSS Japan, Tokyo, Japan).


ECT was performed for a mean of 8.4 ± 2.3 sessions. The severity scores of PID were as follows: score 0 (no delirium), 24 patients; score 1 (mild delirium), eight patients; score 2 (moderate delirium), 10 patients; and score 3 (severe delirium), eight patients. We considered the most severe score for all of the ECT sessions to be the severity score in this study. A PID score >2 was considered to be a clinical problem. Eighteen patients had PID scores of 2–3, therefore the incidence of PID was 36%. In the present study PID was observed 49 times in a total of 419 ECT sessions. A mean of 2.7 ± 2.0 episodes of PID occurred for each patient. Fourteen of 18 patients (77.7%) with PID score 2–3 improved gradually over sequential sessions, and PID was not observed when the catatonic symptoms had ameliorated. In the four remaining patients (22.2%) the occurrence of distinct PID continued until their final sessions. In the group in which PID improved, there were nine patients with schizophrenia and five patients with mood disorder. Nine of 14 patients were responders. In the PID improval group there were six patients with catatonic features before ECT, and catatonic features disappeared gradually in five catatonic patients. In the group in which PID continued, there were three patients with schizophrenia and one patient with mood disorder. Treatment response was observed in two patients. In three out of four patients, catatonic symptoms did not ameliorate until their final sessions.

All PID disappeared within 30 min. Seven of eight patients with severe delirium (score 3) were treated successfully with an i.v. bolus of propofol (1–2 mg/kg). Propofol was used for prevention in only one patient in the fourth session because of extremely severe PID, that is, violent behavior, which was observed from the first to third sessions. PID did not occur in that patient because of the continuous administration of i.v. propofol (4–8 mg/kg per h, 15–20 min) to prevent PID. There was no adverse reaction as a result of the administration of propofol.

Table 1 presents the clinical characteristics between the two groups: patients with no or mild PID (score 0–1) and those with moderate to severe PID (score 2–3). The frequency of patients with catatonic features with moderate to severe PID was significantly higher (P = 0.001) than in those with no or mild PID. There was no difference for the stimulus waveforms (sine or brief pulse square) between the groups with no or mild PID and moderate to severe PID.

Table 1.  Subject data vs severity of PID
 No or mild PID (n = 32)Moderate–severe PID (n = 18)P
  • t-test;

  • Fisher's exact test.

  • ECT, electroconvulsive therapy; PID, post-ictal delirium.

Age (years)50.8 ± 18.743.0 ± 15.00.138
Gender (M/F)8/248/100.211
Duration of illness (years)9.8 ± 10.013.4 ± 13.90.052
Diagnosis  0.051
 Mood disorder166 
 Somatoform disorder50 
Clinical features, n (%)   
 Psychotic features19 (59.4)14 (77.7)0.227
 Catatonic features1 (0.03)8 (44.4)0.001
Stimulus wave form (sine/brief-pulse square)24/812/60.533
No. ECT sessions8.6 ± 2.08.3 ± 2.40.654

Table 2 shows that the occurrence of PID was significantly higher (P = 0.001) in the catatonic group than in the non-catatonic group, and that the other variables were not significantly different between the two groups.

Table 2.  Subject data vs presence of catatonia
 Catatonic (n = 9)Non-catatonic (n = 41)P
  • t-test;

  • Fisher's exact test.

  • ECT, electroconvulsive therapy; PID, post-ictal delirium.

Age (years)49.3 ± 12.447.7 ± 18.80.803
Gender (M/F)4/512/290.442
Duration of illness (years)14.3 ± 10.410.3 ± 11.70.052
Diagnosis  0.518
 Mood disorder418 
 Somatoform disorder05 
Psychotic features before ECT, n (%)7 (77.7)26 (63.4)0.699
Occurrence of PID8 (88.8)10 (24.3)0.001
Stimulus wave form (sine/brief-pulse square)4/532/90.094
No. ECT sessions8.1 ± 2.48.4 ± 2.30.722

Figure 1 shows the incidence of PID, in patients with and without catatonic features. Eight of nine patients with catatonic features had PID (incidence, 88%). Of 41 patients without catatonic features, PID occurred in 10 patients (incidence, 24%). The incidence of PID in catatonic patients was significantly greater than that in non-catatonic patients (P < 0.001). When the data from the patients with schizophrenia and mood disorders without somatoform disorders were reanalyzed, the results were similar to those with somatoform disorder (P =  0.001; Fig. 2).

Figure 1.

Incidence of post-ictal delirium in patients (inline image) with and (□) without catatonic features.

Figure 2.

Incidence of post-ictal delirium, in patients suffering from only mood disorders and schizophrenia (inline image) with and (□) without catatonic features. The data for somatoform disorders were excluded.

Furthermore, the associations between the severity of PID (score 0–3) and the clinical factors are shown in Table 3. Significant correlations were demonstrated for gender (r = 0.163, P < 0.05), diagnosis of schizophrenia (r = 0.352, P < 0.01) and presence of catatonic features (r = 0.482, P < 0.001). Multiple regression, however, showed that scores of PID correlated significantly only with the presence of catatonic features (β = 0.428, P < 0.01).

Table 3.  Severity of post-ictal delirium and clinical factors
  • ***

    P < 0.001,

  • **

    P < 0.01,

  • *

    P < 0.05.

  • r, correlation coefficient of simple regression analysis; β, standardized partial correlation coefficient of multiple regression analysis; R2, coefficient of determination.

Length of illness0.099−0.017
 Mood disorder−0.278−0.160
 Somatoform disorder−0.281−0.139
Clinical features  
 Psychotic features0.209−0.006
 Catatonic features0.482***0.428**
Stimulus wave form−0.250−0.105
R2 0.387*

Sixteen out of 23 schizophrenia patients exhibited a good response (70%) after the ECT sessions. Sixteen out of 22 patients with mood disorder were responsive to the ECT (73%). In all of the patients with somatoform disorder, three patients were responsive to ECT, but the remaining two patients were non-responders. In the present study no relationship was observed between the occurrence of PID and the effects of ECT.


The present study found the incidence of PID to be 36%, while a previous study suggested that the incidence of PID was approximately 10%.6 This difference might be due to different definitions of PID. PID scores of 2–3 were defined as clinically relevant in the present study, whereas the other study evaluated only those in a severe state (i.e. score 3). In addition, we monitored the patients carefully from immediately before ECT to 30 min after the ECT.

The present study showed that PID occurred significantly more often in patients with catatonic features, compared with patients without catatonic features. This is the first report to suggest that catatonic features represent a clinical risk factor for PID. The incidence was extremely high in patients with catatonic features (88%). Although the etiology of PID remains unclear, this finding may be associated with slightly reduced consciousness levels during the immediate post-ictal phase in patients with catatonic features, and the states of PID and the catatonic features may be similar. In fact, most PID improved with remission of the catatonic symptoms.

Differences in the stimulus waveform (sine or brief-pulse square) did not contribute to the appearance of post-ictal delirium. Although anesthesia, electrical stimulation, and seizures each presumably contribute to cause the syndrome, no specific relation to stimulus waveform, dosage, or electrode placement has been confirmed.2 In addition, Devanand et al. failed to identify predictors of PID from among the variables of age, pre-ECT agitation or excitement, number of ECT sessions administered, barbiturate anesthetic, succinylcholine dosage, or mean seizure duration.4 We did not find any association between PID and various factors, including age, gender, duration of illness, diagnosis, psychotic features and stimulus waveform, except for the presence of catatonic features.

In general, PID is readily terminated by i.v. benzodiazepine or barbiturates. Because patients who develop PID manifest this during more than one-third of their treatments, it is standard practice to prevent recurrent episodes by administering diazepam (5–15 mg; or midazolam 1–3 mg) i.v., as soon as the induced seizure terminates. We were able, however, to successfully treat and prevent PID with a propofol bolus (1–2 mg/kg) or infusion (4–8 mg/kg per h, 15–20 min) in all eight patients. Propofol was easily titratable, with predictable effects and a short duration of action, which made the drug easy to handle in the ECT suite. Propofol infusion also allowed precise titration of the sedation to the level of agitation and control of the rate of awakening from the ECT. Emergence from sedation is also rapid due to the fast redistribution of the drug into the peripheral tissues and metabolic clearance.11 There have been some reports on the successful use of propofol to control delirium tremens, which is refractory to benzodiazepines.12–14 In addition, there have been a few case reports on the use of propofol for PID. Augoustides et al. described the successful treatment of severe post-ictal agitation that was resistant to midazolam, lorazepam, haloperidol, with propofol bolus and infusion.15 Sienaert et al. described a case of severe recurrent PID.16 In that study propofol proved to be effective in preventing agitation when it was used as the induction agent or was administered at the end of the seizure, but propofol could not be used to treat a delirious state when it was administered only after the first signs of motor restlessness had emerged. They used a propofol bolus of only 20–40 mg during the emergence phase after ECT.

In conclusion, the present study suggests that the presence of catatonic features before ECT is a predictor of PID. Therefore, preparation for restraint and pharmacological intervention is important in ECT in patients with catatonic symptoms. From this experience we conclude that propofol is a safe and effective agent for the management of PID.