Key words ECT anesthesia - anesthetic change - seizure quality parameters - seizure duration
- ASEI
Introduction
Electroconvulsive therapy (ECT) is an effective treatment for depressive and
psychotic disorders [1 ]
[2 ]
[3 ]. It
relies on the induction of a generalized cerebral seizure under anesthesia and
muscle relaxation. To measure the quality of this seizure, multiple seizure quality
(SQ) parameters have been defined and studied, such as postictal suppression index
(PSI), maximum sustained coherence (MSC), midictal amplitude (miA), average seizure
energy index (ASEI), seizure duration (electroencephalography, EEG/motor),
and peak heart rate (PHR) [4 ]
[5 ]
[6 ]. In
literature, these parameters are generally described as being related to the
therapeutic response [4 ]
[5 ]
[6 ]
[7 ]
[8 ]
[9 ]
[10 ]. Multiple factors influence
the seizure quality, including stimulus dose, electrode placement, patients’
age, and concomitant medication like benzodiazepines [11 ]
[12 ]
[13 ]. Among these factors,
seizure quality is potentially influenced by the anesthetic drug used [14 ]. Until today, there is no (inter-)national
consensus on which drugs should be used for induction and maintenance of general
anesthesia for ECT. The most frequently used anesthetics are barbiturates,
etomidate, propofol, and esketamine [15 ], each
having different characteristics and an impact on seizure quality.
Advantages and disadvantages of different anesthetics used under ECT
Several studies have investigated the effects of different anesthetics on ECT
seizure quality and tolerability. Propofol is an ultrashort-acting anesthetic
but has anticonvulsive characteristics and was found to cause a shortened EEG
seizure [16 ]
[17 ]. (Es-)ketamine is commonly used in ECT
due to its pro-convulsive properties [18 ],
but side effects like nausea, dizziness, and psychotic symptoms appear more
frequently than under other anesthetics during ECT [19 ]. Esketamine – as an enantiomer
of ketamine – has both a higher anesthetic effect and fewer side effects
compared with equally dosed ketamine [20 ].
In spite of a mild anticonvulsive effect, methohexital is a short-acting
barbiturate that has multiple helpful characteristics for use in ECT: it leads
rapidly to a short-lasting narcotic effect, it is not known to have a negative
impact on the length of EEG seizure and has a moderating effect on hemodynamic
parameters like increase of blood pressure or cardiac arrhythmias [14 ]
[17 ]. Under etomidate, myoclonies and a longer wake-up time occur more
often as a side effect than under other anesthetics [21 ]. Besides these findings, a systematic
review of anesthetic agents from 2016 found no difference in the tolerability of
common anesthetics in ECT [22 ]. While it
described ketamine and methohexital to potentially facilitate a higher
antidepressant effect – due to a longer seizure duration than propofol
or thiopental – other reviews and studies could not find differences in
the reduction of depression scores [23 ]
[24 ] despite the inherent
antidepressant effect of ketamine in other treatments. Until now, no study
suggests a significant superiority for one of the mentioned anesthetics or their
combinations.
At the University Medical Center Göttingen, methohexital was used for ECT
anesthesia in most patients until 2019. In 2019 anesthetic drugs in ECT
treatment had to be changed due to unavailability. A combination of
propofol/esketamine was chosen to combine the advantages and reduce the
disadvantages of the two substances as single applications: poor EEG seizures
may improve with lower propofol doses, which can be realized by the addition of
(es-)ketamine [25 ]. Due to renewed
availability, methohexital has been used again since 2022. So far, four studies
compared methohexital with propofol and detected a shorter seizure duration
under the latter [23 ]. Three studies
compared ketamine with methohexital but did not find significant differences in
seizure quality or antidepressant effectiveness [19 ]
[24 ]
[26 ].
To our knowledge, no direct comparisons of methohexital vs. a combination of
propofol/esketamine have been made regarding their effects on SQ
parameters in ECT so far. The current retrospective longitudinal study aims to
close this gap by comparing established SQ parameters before and after the
change from propofol/esketamine to methohexital. To minimize other
factors which may influence the seizure threshold or seizure quality, only
patients undergoing maintenance ECT (mECT) were included. Thus, stimulus dose,
electrode placement, and concomitant medication remained completely stable
throughout the analyzed treatments.
Materials and Methods
Subjects
The following inclusion criteria were applied: (1) patients receiving mECT at our
department irrespective of diagnosis, (2) availability of four consecutive
mECTs, two directly before and two directly after the change from
propofol/esketamine to methohexital, (3) age≥18 years, (4) mECT
within the data collection period from 11/2021 to 04/2022.
We identified 52 patients undergoing mECT, of which 10 were excluded due to
discontinuation of mECT, the necessity of a new ECT series, or intolerability
regarding change of anesthetics. Eight were excluded due to changes in
anesthetics dose, stimulus dose, or electrode placement. Finally, 34 patients
were included in the study (age: 20 to 85 years, means (M )=60.29,
SD= 16.09; 64.7% female), diagnosed with unipolar
depressive disorder (n= 24; ICD-10: F32.2, F32.3 and F33.1 to
F33.4), schizophrenia spectrum (n= 7; ICD-10: F20.0, F20.2,
F25.1), bipolar depressive disorder (n= 2; ICD-10: F31.3 and
F31.8) and dementia with psychotic symptoms (n= 1; ICD-10: F02.8).
All patients had shown a treatment response to the ECT series beforehand. They
received regular mECT for relapse prevention at the Department of Psychiatry and
Psychotherapy, University Medical Center Göttingen.
Concomitant medication (see [Tab. 1 ]) was
kept stable during the course of this study. The study was approved by the local
ethics committee (2/5/22).
Tab. 1 Medication examined in this study.
Antidepressant
Antipsychotic
Anticonvulsants
Lithium
Benzodiazepine
SSNRI
8
–
–
–
–
SSRI
4
–
–
–
–
Tricyclic
1
–
–
–
–
Mirtazapine
2
–
–
–
–
MAO-Inhibitors
2
–
–
–
–
Other
1
–
–
–
–
Combination
12
–
–
–
–
None
4
–
–
–
–
Atypical
–
17
–
–
–
Combination
–
8
–
–
–
None
–
9
–
–
–
Pregabalin
–
–
2
–
–
Lamotrigine
–
–
1
–
–
None
–
–
31
–
–
Lithium
–
–
–
8
–
None
–
–
–
26
–
Benzodiazepine
–
–
–
–
5
None
–
–
–
–
29
Notes. Medication for N =34 patients; SSNRI,
selective serotonin norepinephrine reuptake inhibitors; SSRI, selective
serotonin reuptake inhibitor; MAO-Inhibitors, monoamine oxidase
inhibitors.
Study Design
For each patient, data from four mECT treatments was gathered, pre-
(T
1
and T
2
) and
post-change (T
3
and T
4
) from
propofol/esketamine to methohexital (see above). A total of eight
established SQ parameters (see [Tab. 2 ])
were measured: (1) PSI, (2) ASEI, (3) MSC, (4) miA, (5/6) seizure
duration (motor, cuff method), and EEG, (7) seizure concordance T
1
: n= 33,
T
2
: N= 34,
T
3
: n= 33,
T
4
: n= 31).
Tab. 2 Definition of SQ parameters.
1. Postictal suppression index (PSI)
Measures the decrease of the EEG amplitude directly at the
end of the seizure in %
2. Average seizure energy index (ASEI)
Is the integral of the seizure amplitude over time divided by
the duration of the seizure
3. Maximum sustained coherence (MSC)
Measures the synchronization of convulsions between the
hemispheres in %
4. Midictal amplitude (miA)
Describes the maximal ictal amplitude in a seizure in
µV
5. Motor seizure duration
Is defined by the length of motoric convulsions, here
measured by the cuff method in seconds
6. EEG seizure duration
Is measured by EEG and shows the total length of the seizure
in seconds
7. Seizure concordance
Calculates the concordance between motor and EEG seizure
8. Peak heart rate (PHR)
Describes the maximum heart rate during the seizure, measured
in beats per minute
Notes. Source: Instruction manual from the Thymatron IV device
(Somatics, LLC., Lake Bluff, IL, USA; [44]); EEG,
electroencephalography; SQ, seizure quality.
Maintenance (m-) electroconvulsive therapy
MECT was performed with a Thymatron IV device (Somatics, LLC., Lake Bluff, IL,
USA). The double-dose program and brief pulse technique were used (maximum dose
of 1008mC; 200%). Initially, the stimulus dose for the first ECT
treatments was determined age-based. Both dosing and electrode placement had
been previously adjusted depending on clinical response, tolerability, and
seizure quality during acute ECT. To eliminate potential intra-individual
confounding, only patients with constant stimulus dose and electrode placement
were included. All patients received a combination of propofol/esketamine with
constant dosage for the first two sessions (T
1
and
T
2
). In most of the cases, the proportion of
propofol was higher (M =0.88 mg/kg) when compared to
esketamine (M =0.68 mg/kg). The dosages had been initially
adjusted over the course of treatment before mECT: At the beginning, most
patients had received a dosage of 1 mg propofol/kg body weight and
0.5 mg esketamine/kg body weight.
For the second two mECT sessions (T
3
and
T
4
), all patients received a constant dosage of
methohexital. Here, the initial dosage before mECT was 1 mg
methohexital/kg body weight, and M= 1.21 mg/kg during
mECT.
Statistical analyses
IBM SPSS Statistics 29 (IBM Corp. Armonk, NY) was used to analyze data. For
descriptive representation, means (M ) and standard deviations (SD) were
created for numeric variables, as well as Pearson correlations (r ). To
analyze the main outcome (change of SQ parameters), eight general linear models
(GLM) for repeated measures were used. Measurements were included as a
four-staged within-subjects factor (mECT sessions: T
1
to T
4
). Pairwise comparisons could be calculated both
within a constant condition of anesthesia (propofol/esketamine:
T
1
vs. T
2
;
methohexital: T
3
vs. T
4
) and
between two conditions of anesthesia
(T
1
/T
2
vs.
T
3
/T
4
), enabling us to
control for intrapersonal variations independently of anesthesia changes. For
multiple comparisons, p -values were corrected within each model
(Bonferroni method; initial significance at p <0.05 before
correction, two-tailed). Exploratory models controlled for age (see results
section for details). For all SQ parameters except for PSI (see above),
N= 34 patients provided complete datasets.
Results
Descriptive results
Please see [Tab. 3 ] for an overview.
Electrodes were placed left anterior right temporal (n= 15), right
unilateral (n= 10), and bitemporal (n= 9). The mean
stimulus dose was M= 109.12% (SD= 54.57;
100%=504mC). The mean PSI in percent reached
M= 75.59% (SD= 15.10), ASEI was
M= 13.14 = 9.06), MSC (0% to 100%) was
M =95.71% (SD =5.63), and midictal amplitude
was M= 190.25 (SD=57.65). The patients showed
M= 35.51 s (SD=12.93) motor seizure duration
and M= 52.26 s (SD=17.33) EEG seizure duration.
Seizure concordance was M= 69.12% (SD=15.79). The
mean peak heart rate was M= 126.24 beats/minute
(SD= 19.08). Significant correlations were found between SQ
parameters (see [Tab. 3 ]; variables 4 to
11). A higher stimulus dose was applied to older patients
(r= 0.458, p< 0.01). Both increasing age and higher
stimulus dose were negatively correlated with SQ parameters (age: r
between -0.022 and -0.514, p< 0.05/0.01 in 5 out of 8 SQ
parameters; stimulus dose: r between -0.190 and -0.668,
p< 0.05/.01 in 7 out of 8 SQ parameters).
Tab. 3 Correlations, means, standard deviations, and
frequencies.
Variable
1
2
3
4
5
6
7
8
9
10
M±SD/Freq.
1. Age
–
60.29±16.09
2. Gender
–.153
–
f: 22, m: 12
3. Stim. dose
.458**
–.172
–
109.12±54.57
4. PSI
–.514**
.325
–.412*
–
75.59±15.10
5. ASEI
–.503**
.128
.637**
.484**
–
13.14±9.06
6. MSC
–.119
–.018
–.384*
.027
.096
–
95.71±5.63
7. Midictal amplitude
–.468**
.125
–.668**
.478**
.966**
.167
–
190.25±57.65
8. Seiz. dur. (motor)
–.240
.224
–.409*
.355*
.067
.458**
.059
–
35.51±12.93
9. Seiz. dur. (EEG)
–.022
.006
–.190
–.039
–.221
.370*
–.222
.840**
–
52.26±17.33
10. Seiz. concordance
–.379*
.351*
–.372*
.721**
.485**
.237
.467**
.430*
–.103
–
69.12±15.79
11. Peak heart rate
–.438**
.336
–.458**
.421*
.211
.443**
.234
.521**
.359*
.372*
126.24±19.08
Notes. *p <.05.
**p <.01. Captions: Gender
(f=female/1, m=male/2); stimulus
dose (0% to 200%; 100%=504mC);
postictal suppression index (PSI; 0% to
100%); ASEI (seizure
duration in seconds (motor and EEG); seizure concordance
maximum sustained coherence (MSC; 0% to
100%); midictal amplitude (µV)
(N= 34); PSI, postictal suppression index; ASEI, average
seizure energy index; MSC, maximum sustained coherence; EEG,
electroencephalography; Seiz. dur. seizure duration.
For the first two mECT sessions (T
1
and
T
2
), the mean dosage of propofol was
M= 64.26 mg (SD=20.04,
min=40 mg, max=120 mg,
M= 0.88 mg/kg), and the mean dosage of esketamine was
M= 51.18 mg (SD=16.65,
min=20 mg, max=80 mg,
M= 0.68 mg/kg). For the second two mECT sessions
(T
3
and T
4
), the mean
dosage of methohexital was M= 88.97 mg (SD=22.99,
min=50 mg, max=140 mg,
M= 1.21 mg/kg). The mean interval between the two mECTs
was M= 4.28 weeks (SD= 2.44). After the change to
methohexital, two patients required prophylactic antiemetic medication and one
patient needed prophylactic analgesic medication to prevent headaches in further
treatments.
Longitudinal analysis of seizure quality parameters
[Fig. 1 ] and [Fig. 2 ] present graphical summaries of all
SQ parameters. The (1) PSI did not vary significantly between the four mECT
sessions (F (3, 87)=2.29, p= 0.084, partial
η2 =0.07; see [Fig. 1a ]). For the (2) ASEI, a general variation was found
(F (3, 99)=4.66, p= 0.004, partial
η2 =0.12; see [Fig. 1b ]). Corrected pairwise comparisons showed a significant
decline of the ASEI from T
2
(M= 11.56)
compared to T
3
(M= 7.75,
p= 0.039) and T
4
(M= 6.12, p= 0.013). Numerically, the difference
between T
1
and
T
3
/T
4
was even
higher but did not reach significance due to a higher variance at
T
1
(p= 0.181/0.106; see
[Fig. 1b ]). The (3) MSC did not vary
significantly between the ECT sessions (F (3, 99)=0.338,
p= 0.80, partial η2 =0.01; see [Fig. 1c) ], in contrast to the (4) midictal
amplitude (F (3, 99)=8.52, p< 0.001, partial
η2 =0.01; see [Fig. 1d ]): Pairwise comparisons showed a significant decline from
T
1
(M= 214.28)/T
2
(M= 205.05) to T
3
(M= 172.25) and T
4
(M= 169.42; p= 0.022 to<0.001). In sum, a
significant decrease with the use of methohexital could be found exclusively for
ASEI and midictal amplitude.
Fig. 1 Course of seizure quality parameters in psychiatric
patients during four maintenance electroconvulsive therapy (mECT)
sessions, pre- (T
1
/T
2
)
and post- (T
3
/T
4
)
anesthesia change. p< 0.05 *,
p< 0.01**,
p< 0.001***. Mean values with
95%-CIs and Bonferroni corrected pairwise comparisons;
(a) postictal suppression index (PSI); (b) average
seizure energy index (ASEI); (c) maximum sustained coherence
(MSC); (d) midictal amplitude.
Fig. 2 Course of seizure quality parameters in psychiatric
patients during four maintenance electroconvulsive therapy (mECT)
sessions, pre- (T
1
/T
2
)
and post- (T
3
/T
4
)
anesthesia change. p< 0.05*,
p< 0.01**,
p< 0.001***. Mean values with
95%-CIs and Bonferroni corrected pairwise comparisons;
(e) motor seizure duration; (f) seizure duration
electroencephalography; (g) seizure concordance ((h)
peak heart rate (PHR).
Seizure duration varied significantly between measurements, both for (5) motor
(F (3, 99)=13.90, p< 0.001, partial
η2 =0.30; see [Fig. 2e ]) and (6) EEG (F (3, 99)=22.11,
p< .001, partial η2 =0.40; see [Fig. 2f ]). Motor seizure duration raised
from T
1
(M= 30.76)/T
2
(M= 30.35) to T
3
(M= 39.26) and T
4
(M= 41.65; p= 0.012 to<0.001). Likewise, the
EEG seizure duration raised from T
1
(M= 44.29)/T
2
(M= 44.00) to T
3
(M= 59.15) and T
4
(M= 61.62; all p <0.001). Seizure concordance (7)
did not vary significantly between the measurements (F (3,
99)=0.60, p= 0.615, partial
η2 =0.02; see [Fig.
2g ]). There was no significant variation for the (8) PHR between
measurements (F (3, 99)=0.367, p= 0.777, partial
η2 =0.01; see [Fig. 2h ]). In sum, seizure duration was significantly longer with the
use of methohexital, both for motor (approx.+10 s) and for EEG
(approx.+15 s).
Influence of age on seizure quality parameters
Numerous studies have found that elderly patients show inferior SQ parameters
[11 ]
[12 ]
[13 ], which may lead to
treating them with higher stimulus doses (see [Tab. 3 ]). We created two subgroups of patients based on the median
age (63+years vs.<63 years; each group n= 17). A
two-staged between-subjects factor was then added to each of the eight GLMs
reported above to analyze general differences between both groups (main effect:
between groups) or different possible trajectories between older vs. younger
patients depending on the anesthetic used (interaction effect). In sum, we did
not find a significant effect between both groups (p= 0.094 to
0.619) or an interaction effect (p= 0.176 to 0.771) for any GLM.
Numerically, older patients showed worse SQ parameters, but differences were too
small to reach significance and remained constant over the anesthesia changes
for all parameters.
Discussion
In this retrospective study, we longitudinally analyzed the influence of different
anesthetics on SQ parameters. Therefore, data from four mECTs was gathered, pre- and
post-change from propofol/esketamine to methohexital for each patient. Both
ASEI and midictal amplitude showed a significant decrease under methohexital if
compared to propofol/esketamine, whereas seizure duration (motor and EEG)
was significantly longer under methohexital. PSI, MSC, seizure concordance, and PHR
remained stable.
Before interpreting these findings, it must be noted that uncertainties do still
exist regarding SQ parameters. Although the quality of a seizure under ECT is
evaluated on the basis of wave amplitudes, seizure duration, PSI, MSC, and PHR [5 ]
[9 ]
[27 ], to what extent these
parameters correlate with the therapeutic efficacy remains ambiguous. For example,
elderly patients have higher seizure thresholds but respond more often to ECT [28 ]
[29 ],
and though benzodiazepines may decrease SQ parameters, their use does not seem to
reduce the effectiveness of ECT [30 ]. Also,
factors associated with better SQ parameters, like younger age or hyperventilation
right before ECT, do not necessarily result in a better therapeutic effect [28 ]
[31 ].
Furthermore, clinical predictors improving the probability of ECT response like
psychotic/catatonic symptoms, fewer previous medication failures, short
illness episodes or absence of comorbid personality disorder do not influence
seizure quality markers [32 ]
[33 ]
[34 ].
Nonetheless, these parameters represent the best-researched predictors for ECT
effectiveness to date [4 ].
In this study, we found that seizure duration heavily depended on the choice of
(combined) anesthetic substances. There is some evidence that notably short seizure
duration (depending on the source less than 15 or 24 s) is leading to a
poorer clinical outcome [5 ]
[35 ] – this would argue for the use of
methohexital over propofol/esketamine in cases of borderline seizure
duration. However, other studies did not show a significant correlation between
seizure duration and clinical outcome [9 ]
[36 ]
[37 ].
Regarding the meaning of ASEI and midictal amplitude for the therapeutic outcome,
different studies come to heterogeneous results: some outpoint a correlation between
a decrease in depressive symptoms and a higher wave amplitude [4 ]
[5 ]
[27 ], others do not [7 ]
[10 ].
Furthermore, as described above, there is a negative correlation between wave
amplitude, seizure length, and age [11 ]
[38 ]
[39 ],
with age being considered a positive predictor for ECT response [28 ]
[34 ]
[40 ]. Our results also show that
some of the established SQ parameters listed above (PSI, MSC, seizure concordance,
PHR) are not influenced by the change of anesthetics. So far, mostly PSI [7 ]
[10 ]
[27 ] and MSC [7 ]
[27 ]
have been positively associated with a better therapeutic effect.
In conclusion, this study clearly shows differential effects of the anesthetics
methohexital vs. propofol/esketamine on four out of eight analyzed SQ
parameters: methohexital is associated with a longer seizure duration, whereas
propofol/esketamine lead to higher amplitudes. However, it is not possible
at this time to make a definitive statement about their relationship – or
the relationship of propofol/esketamine vs. methohexital – to
treatment response.
From a pharmacological point of view, both the combination of
propofol/esketamine as well as the use of methohexital are suitable
approaches for inducing general anesthesia for ECT. Both substances show a very
quick onset of 10–30 s after infusion with a duration of action of
5–10 min, which makes them suitable for short-lasting procedures.
The effects on the central nervous system significantly differ between the two
substances. Methohexital application initially leads to biphasic EEG changes with
the occurrence of excitatory, proconvulsive symptoms, especially in low to moderate
doses. A state of burst-suppression is reached only after high doses. Propofol has
a
dose-dependent anticonvulsive effect even in low doses, which can be a limitation
for its use in ECT patients. It is, therefore, usually combined with a second
hypnotic substance such as ketamine, to avoid relevant anticonvulsive concentrations
altering ECT quality. Ketamine leads to dissociative anesthesia. It is not known to
have relevant anticonvulsive effects.
Regarding the tolerability of propofol/esketamine vs. methohexital, we
examined concomitant medication of patients during treatment and found that in two
cases, a new medication was started to prevent (1) nausea and (2) headaches after
switching to methohexital. Whereas barbiturates are known to cause postanesthetic
nausea and vomiting in a relevant proportion of patients, propofol is known to
prevent these side effects [41 ]
[42 ]. As numbers are very small and a direct
assessment of symptoms in the patients was missing, so rather mild symptoms might
have been overlooked, and at this point, no general statement can be made.
Therefore, a future prospective design with a focus on tolerability (and possibly
treatment response) would be necessary. In summary, neither methohexital nor
propofol/esketamine was clearly superior regarding the influence on seizure
parameters, tolerability or clinical applicability.
Limitations and strengths
There are some limitations regarding this study. First, as the study relied on
retrospective examination of longitudinal clinical data, possible effects of the
different anesthetics on tolerability were not systematically examined. As
discussed above, this would facilitate implications for clinical routine and
should be focused on in the future, then prospective studies. Second, due to the
retrospective design, dosages of anesthetics had been chosen according to
clinical standard but without a consistent dosing protocol. Therefore, it cannot
be stated with absolute certainty that the depth of anesthesia was equivalent
between subjects and may have impacted seizure quality. The same applies to the
time interval between anesthetic administration and stimulation: this had also
been done according to clinical standard, but timings had not been protocolled
during clinical routine and could thus not be analyzed within the framework of
this retrospective study. Furthermore, it was not possible to correlate
anesthesia with therapeutic implementations, which should be considered in a
future prospective design. Third, the generalization of results is limited due
to our relatively small sample size, although our sample represented a typical
set of stable treated mECT patients. Only larger and prospective, comparative
trials focusing on acute ECT could help to ultimately clarify the differential
effects of different anesthetic regimens regarding (a) SQ parameters, (b)
effectiveness, and (c) tolerability. This would also allow the addition of
separate samples, in each of which only propofol or esketamine could be
administered as the sole anesthetic, to analyze differential effects. However,
we would like to point out that the longitudinal study design presented here
largely eliminated interference factors (e. g., changes in stimulus
dose, electrode placement, etc.), which is a clear strength of this study.
Ethical approval
This study has been approved by the local ethics committee and has therefore been
performed in accordance with the ethical standards laid down in the 1964 Declaration
of Helsinki and its later amendments.
