Remission rate of transcranial magnetic stimulation compared with electroconvulsive therapy: a case-control study

Abstract

Purpose: To compare the rate of remission, rate of response, change in depressive symptoms, and adverse effects between repetitive transcranial magnetic stimulation (rTMS) and electroconvulsive therapy (ECT).

Materials and methods: In this retrospective case-control study, 35 patients treated for depression with rTMS (left dorsolateral prefrontal cortex, 90% observed motor threshold, 10 Hz, 2000 pulses/session, 15 sessions) at Örebro University Hospital, Sweden (cases), were compared with a matched group of 35 patients treated for depression with ECT (controls). Data on controls were obtained from the Swedish National Quality Register for ECT (Q-ECT). Severity of depression was evaluated using the Montgomery-Åsberg Depression rating scale (MADRS).

Results: Remission rate was 26% for cases and 43% for controls (p = .3). Response rate was 40% for cases and 51% for controls (p = .63). The median decrease in MADRS was 11 (IQR 3–19) vs. 17 (IQR 6–27; p = .10) for rTMS and ECT, respectively. There was no statistically significant difference in any measure of treatment effect between rTMS and ECT. More than half of the patients of the rTMS group experienced scalp discomfort and 11% of the ECT group had memory disturbances.

Conclusions: All measures of therapeutic efficacy were numerically inferior in the rTMS group compared to the ECT group. The differences were not statistically significant, probably because the sample size was small. More studies are required to find the optimal place for rTMS within the Swedish health care system. Such studies could be facilitated by inclusion of rTMS in the Q-ECT.

Keywords:

• Major depressive disorder
• repetitive transcranial magnetic stimulation
• electroconvulsive therapy
• treatment outcome

Background

Major depressive disorder (MDD) impairs social functioning, causes personal suffering, and has economic consequences for the individual and society due to long-term incapacitation and loss of production [1]. The 12 month prevalence of MDD is about 5%. The symptoms can differ between patients and 30–60% of all cases are not improved significantly by antidepressant medication [2,3]. These cases are considered to have treatment-resistant depression (TRD) [4].

An emerging treatment for TRD is repetitive transcranial magnetic stimulation (rTMS) [5]. This is a noninvasive technique that utilizes pulsating shifting magnetic fields generated by extracorporeal electromagnetic coils to focally excite or inhibit cortical neurons through electromagnetic induction [6]. Although the therapeutic mechanism of rTMS remains unclear, it may involve changes in neuroplasticity due to long-term potentiation, long-term depression, and dendritic growth owing to increases in neurotrophic factors [7,8]. There is evidence that rTMS affects depressive symptoms [9–12], but it is unclear if the effects are large enough to be clinically relevant [13,14]. The therapeutic effects correlate with the number of pulses and number of sessions in some studies [10–12,15,16] but not others [17]. Patients often have more than one concommittant treatment [18]. There is some evidence of a synergistic antidepressant effect of rTMS with antidepressant medication [12]. In one study, younger age predicted the efficacy of rTMS for patients with unipolar depression but not bipolar depression [19]. Other reported predictors of efficacy are short duration of depressive symptoms [19,20] and cognitive rather than somatic depressive symptoms [21].

Electroconvulsive therapy (ECT) is an established treatment for MDD, particularly TRD [22]. In ECT, generalized seizures are induced under general anesthesia by stimulation with electrical current [22]. The efficacy of rTMS for depression is inferior to that of ECT [23]. ECT seems especially preferable to rTMS for treatment of MDD with psychotic symptoms [23,24], while for non-delusional MDD rTMS may have similar effect [23]. ECT, but not rTMS, is associated with transient memory loss [25].

In 2017, the Swedish National Board of Health and Welfare issued national recommendations for the use of rTMS for the treatment of MDD [26]. These were based on results of international studies and include rTMS as a treatment that ‘can be offered’ for moderate to severe MDD. There are no data on the treatment results of rTMS in Swedish psychiatry and there is a need for scientific evaluation of this novel approach and its role in the treatment of depression.

Aim: To establish the rate of remission, response, change in depressive symptoms, and adverse effects with rTMS and to compare it with ECT when used for the treatment of depression in a Swedish population.

Materials and methods

Study design

This was a retrospective case-control study of 35 patients who were treated with rTMS at the University Hospital Örebro, Sweden, between 2003 and 2007. Controls (the ECT group) were 35 patients treated with ECT in Sweden between 2012 and 2016 and were reported to the the Swedish national quality register for ECT (Q-ECT) [27]. rTMS and ECT groups were matched for age, gender, diagnosis, psychotic symptoms, and depression severity, evaluated using the Clinical Global Impressions Severity [28] scale. All cases and controls had TRD of at least stage 1 on the Thase and Rush scale [4], i.e. failure to achieve response after a trial with antidepressant medication in adequate dose and duration.

Patient selection and data sources

Data on cases were obtained from medical records. There were 42 medical records of rTMS treatment series within the identified time period, distributed over 37 patients. Five patients had more than one rTMS treatment series on record, and all but the first treatment series were discarded. Two patients were excluded because treatment was stopped prematurely due to overheating of the magnetic coil causing machine malfunction. Data from the remaining 35 patients were included in the analysis (Figure 1). Data on controls were obtained from the (Q-ECT) [27]. Thirty-five ECT-treated patients were matched for gender, age, diagnosis, depression severity (Clinical Global Impressions Severity score), and psychotic symptoms, in that order, to the rTMS population (Figure 1). The ECT patients were registered from 19 different psychiatric clinics in Sweden. Information about prior personality disorder was collected from the Swedish national patient register, which includes diagnoses from more than 99% of hospital discharges. Information about prior antidepressant treatment was collected from the Swedish prescribed drug register, which has a coverage of 100% of medication dispensed through pharmacies.

Figure 1. Flow chart of patient selection. Forty-two rTMS treatments were found, of which thirty-five were unique first-time rTMS treatments. Two rTMS patients were excluded due to technical malfunction. Thirty-five rTMS patients was included in the study and were matched to thirty-five ECT patients from the Swedish National Quality Register for ECT to function as control group.

Repetitive transcranial magnetic stimulation (rTMS)

All rTMS was performed with the Magventure MagPro R30 stimulator and the Magventure MCF-75 circular coil, both manufactured by Magventure A/S Farum, Denmark. The coil was first positioned at 90°over the central sulcus on the patient’s scalp and then shifted to a 45 degree angle to find the left motor cortex. The coil was moved until stimulation induced a movement of the right index finger. The coil was then kept in this location and the stimulation intensity was gradually reduced until no visible movement of the index finger was observed. The stimulation intensity at this point was defined as the observed motor threshold (OMT). After OMT was established, the coil was moved 5 cm anterior so that it was over the left dorsolateral prefrontal cortex and the stimulation intensity was set to 90% of OMT. rTMS was then performed, consisting of 2000 pulses divided into 20 trains of 100 pulses at a frequency of 10 Hz. A 20 s inter-train pause allowed for cooling of the coil. A treatment series consisted of 15 sessions and was not adjusted according to treatment response.

Electroconvulsive therapy (ECT)

ECT is usually administered thrice weekly under general anesthesia. Patients were sedated with intravenous propofol or thiopental, succinylcholine was administered as a muscle relaxant, and atropine or another anticholinergic drug was administered to alleviate the risk of cardiac arrhythmias during stimulation when indicated. Of the 35 ECT patients included in this study, 75% had unilateral and 25% had bilateral electrode placement. The median number of sessions were six (inter quartile range 6–10). ECT was delivered with the following parameters: charge, 318 ± 132 mC; current, 845 ± 55 mA; duration, 69 ± 12 s; frequency, 53 ± 21 Hz; pulse width, 50 ± 6.8 ms; seizure time, 41 ± 25 s (all values are mean ± standard deviation). Detailed information about the ECT of each patient is available in Supplementary Table 1.

Table 1. Baseline characteristics.

	rTMS n = 35	ECT n = 35	p Value
Age (years)	46 (32–57)^a	46 (33–62)^a	.89^b
Gender			1.00^c
Male	74%	74%
Female	26%	26%
Treatment indication			1.00^c
Unipolar depression	71%	71%
Bipolar depression	26%	26%
Schizoaffective disorder	3%	3%
Psychotic symptoms	23%	23%	1.00^c
History of ECT	43%	74%	.003^c
In-patient care during treatment	23%	91%	<.001^c
Concommittant antidepressant	89%	80%	.46^c
Concommittant antipsychotic	46%	51%	.80^c
Concommittant mood-stabilizer	34%	43%	.47^c
Personality disorder	26%	20%	.41^c
CGI-S score	4 (4–5)^a	4 (4–5)^a	1.00^a
MADRS score	30 (23–36)^a	34 (25–41)^a	.14^b

CGI-S: Clinical Global Impressions Severity scale; ECT: electroconvulsive therapy; MADRS: Montgomery-Åsberg Depression Rating Scale or Montgomery-Åsberg Depression Rating Scale self-assessment.

^a Median (interquartile range).

^b Mann-Whitney U-test.

^c Fisher’s exact test.

Outcome measures

In the rTMS group, the severity of depression was evaluated using the Montgomery-Åsberg Depression Rating scale (MADRS) [29] at baseline (pre-treatment), after seven sessions of rTMS, at the final session of rTMS, and at follow-up (45 days after the first session of rTMS). In the ECT group, the severity of depression was evaluated using MADRS or the self-rated version of the MADRS scale (MADRS-S) [30], before ECT and within one week after the last session. Remission was defined as MADRS or MADRS-S score ≤10 after last stimulation. Response was defined as a 50% decrease in MADRS or MADRS-S score after last stimulation compared with baseline. For each rTMS session, the treating clinician used a questionnaire to document patient fatigue, headache, subjective flushing, discomfort, and epileptic seizures. Each item was rated as either present or absent. These data were used to evaluate adverse effects in the rTMS group. In the ECT-group, subjective memory dysfunction was evaluated using a clinician rated scale ranging from zero (no memory dysfunction) to six (total inability to remember). This was combined with clinical records of clinician assessed adverse effects available in the Q-ECT.

Statistical analysis

The primary outcome was remission. Secondary outcomes were response and change in MADRS or MADRS-S score (ΔMADRS) from baseline to last treatment. Outcomes were compared across rTMS and ECT groups using Fisher’s exact test, due to the presence of cells with values <5 within nominal variables and the Mann-Whitney U-test for interval and ordinal non-normally distributed variables. The significance level was p < .05. In two separate subgroup analyses, cases and controls were compared after stratification for age (<50 and ≥50 y) and the presence or absence of psychotic symptoms. The cutoff for the age-stratified analysis was chosen as 50 y in accordance with evidence of decreased efficacy of rTMS in patients ≥50 y [18]. Patients for whom the post-treatment MADRS score was missing were included in the post-treatment analyses with the patient’s nearest historical assessment scale score (last observation carried forward). If there was no post-baseline assessment, drop-outs were defined as non-remissive.

Ethics

All cases consented to rTMS treatment and detailed clinical follow-up to facilitate evaluation of the treatment and were not required to provide additional consent for this chart review. Data from controls were obtained from the Q-ECT. Participation in this register is voluntary and at the time of entry to the register patients were informed of future scientific use of their data. The study was approved by the regional ethics committee in Uppsala (approval no. 2016/196).

Results

Included patients

Seventy patients were included in the study: 35 cases who received rTMS and 35 controls who received ECT. There was a larger proportion of in-patients among controls than cases, and a history of ECT treatment was more common in controls than cases (Table 1). There were no other statistically significant differences between groups among measured variables (Tables 1 and 2).

Table 2. Classes of concommitant antidepressants used.

Antidepressant	rTMS	ECT
SSRI	20%	63%
SNRI	29%	43%
TCA	29%	11%
Mirtazapin	23%	34%

SSRI: selective serotonin reuptake inhibitor; SNRI: serotonin noradrenaline reuptake inhibitor: TCA: tricyclic antidepressant.

Remission and response rate

The rate of remission was 26% in cases and 43% in controls (p = .3). The rate of response was 40% in cases and 51% in controls (p = .63). Median ΔMADRS was 11 in cases and 17 in controls (p = .10; Table 3).

Table 3. Treatment details and results.^e

	rTMS n = 35	ECT n = 35	p Value
Number of sessions	15 (15–15)^a	6 (6–10)^a
Treatment duration (days)	20 (18–21)^a	14 (11–21)^a
MADRS score (last treatment)	16 (8–28)^a	13 (8–28)^a	.35^b
ΔMADRS	11 (3–19)^a	17 (6–27)^a	.10^b
MADRS response0^d	14 (40%)	18 (51%)	.63^c
MADRS remission0^d	9 (26%)	15 (43%)	.30^c

MADRS: Montgomery-Åsberg Depression Rating Scale or Montgomery-Åsberg Depression Rating Scale self-assessment.

^a Median (interquartile range).

^b Mann–Whitney U-test.

^c Fisher’s exact test.

^d Percentages within treatment group.

^e Post-treatment MADRS score was missing for three rTMS patients; therefore, nearest historic value carried forward for MADRS score, ΔMADRS, MADRS response, and MADRS remission for these patients.

Three cases dropped out during treatment and did not complete all MADRS assessments: one relapsed into alcohol abuse after six sessions of rTMS, one opted out of treatment due to lack of effect after nine sessions of rTMS, and one was admitted to emergency psychiatric care after an attempted suicide. Four controls received fewer than six sessions of ECT. These four individuals were non-remissive and non-responding, reasons for terminating treatment was not reported.

Subgroup analyses

Remission, response and median ΔMADRS in rTMS and ECT groups after stratification according to age or the presence or absence of psychotic symptoms are presented in Tables 4 and 5. There were no statistically significant differences but the ECT group had numerically superior results in all subgroups.

Table 4. Treatment effect in patients stratified according to age.^e

	Age <50 years			Age ≥50 years
	rTMSn = 20	ECT n = 20	p Value	rTMS n = 15	ECT n = 15	p Value
ΔMADRS	15 (6–25)^a	19 (5–28)^a	.76^b	9 (7–24)^a	13 (7–24)^a	.28^b
MADRS response^d	8 (40%)	10 (50%)	1.0^c	6 (40%)	8 (53%)	.47^c
MADRS^d remission	4 (20%)	7 (35%)	.72^c	5 (33%)	8 (53%)	.27^c

MADRS: Montgomery-Åsberg Depression Rating Scale or Montgomery-Åsberg Depression Rating Scale self-assessment.

^a Median (interquartile range).

^b Mann-Whitney U-test.

^c Fisher’s exact test.

^d Percentages within stratified treatment group.

^e Comparison of the measures between rTMS and ECT within groups stratified according to age. MADRS score was missing for three rTMS patients; therefore, nearest historic value carried forward for MADRS score, ΔMADRS, MADRS response, and MADRS remission for these patients.

Table 5. Treatment effect in psychotic and non-psychotic patients.^e

	Psychotic			Non-psychotic
	rTMS n = 8	ECT n = 8	p Value	rTMS n = 27	ECT n = 27	p Value
ΔMADRS	16 (4–25)^a	23 (11–28)^a	.54^b	14 (4–21)^a	16 (5–27)^a	.5^b
MADRS response^d	4 (50%)	5 (63%)	.38^c	10 (37%)	13 (48%)	1.0^c
MADRS remission^d	3 (38%)	3 (38%)	.62^c	6 (22%)	12 (44%)	.25^c

MADRS: Montgomery-Åsberg Depression Rating Scale or Montgomery-Åsberg Depression Rating Scale Self-assessment.

^a Median (interquartile range).

^b Mann-Whitney U-test.

^c Fisher’s exact test.

^d Percentages within stratified treatment group.

^e Comparison of the measures between rTMS and ECT within groups stratified according to presence or absence of psychotic symptoms. MADRS score was missing for three rTMS patients; therefore, nearest historic value carried forward for MADRS score, ΔMADRS, MADRS response, and MADRS remission for these patients.

Adverse effects

Twenty-five cases (71%) reported at least one adverse effect in at least one rTMS session. Twenty cases (57%) reported scalp discomfort, 18 (51%) reported headache, 14 (40%) reported fatigue, and five (14%) reported flushing. No rTMS patient had an epileptic seizure. In the ECT population, 11% reported impairing or embarrassing memory dysfunction.

Discussion

The rates of remission and response were lower in the rTMS-group as compared to the ECT group (26 vs. 43% and 40 vs. 51%, respectively), but these differences were not statistically significant. Subgroup analyses revealed similar results across subgroups after stratification according to age (<50 and ≥50 years) or psychotic symptoms. Our results are in line with previous studies showing rTMS to be significantly inferior to ECT for treatment of depression [23]. The likely explanation for insignificant differences in this study is the small sample size examined.

Regarding the technical aspect of the rTMS treatment relatively low stimulus intensity (90% MT) was chosen to limit the risk for epileptic seizures. The amount of pulses (2000 per session) was higher than seems to be optimal [10], which seems to be 1200–1500 pulses per stimulation. This may have limited the therapeutic effect. The determination of OMT and the position of the coil used in the treatment of cases included in this study was imprecise compared with that achieved with neurographic methods [31] and does not account for differences in head size, skull asymmetry, or individual neuroanatomy. Correct positioning of the coil is as important as stimulation intensity and the efficacy of rTMS may be enhanced by utilizing neuroimaging-guided positioning [17]. With precise positioning of the coil, rTMS at an intensity of 90% motor threshold (MT) appears to have therapeutic effect [17,24]. However, with imprecise positioning of the coil, even stimulation as high as 120% MT may be insufficient and there does not appear to be any increased therapeutic effect with stimulation intensities between 93 and 120% MT [17]. However, the method of determining OMT used for patients in the current study may result in an overestimation of the required stimulation intensity by as much as 28% of maximum stimulator output [31]. As such, it is uncertain if there was sufficient stimulation of the dorsolateral prefrontal cortex in all patients.

Comparing rTMS results across studies is complicated because of heterogenic rTMS protocols [32]. There are also different coils being used [33]. The pulse frequency and location used in this study also have heterogenic evidence of efficacy, with publications of both statistically significant therapeutic effect [34], and of no significant therapeutic effect compared to sham-rTMS [35,36]. Nevertheless, the rate of remission in the patients in this study is comparable to that reported in international studies comparing rTMS to sham [9,24,34,36,37].

The response rate with ECT (51%) was not optimal in the current study, compared to 80% in a broader sample from the Q-ECT [38]. Technical aspects of ECT that could be improved for the ECT patients included in this analysis are longer pulse width, more frequent use of bilateral electrode placement, less use of benzodiazepines, and for some patients, longer treatment series [39]. For example, if a patient does not respond to unilateral stimulation after six or more sessions, they could be switched to bilateral stimulation to potentiate efficacy. This was not done for any of the controls in the current study.

Limitations

The small sample size makes the comparisons uncertain, especially among subgroups. A potential confounder in the present analysis is the larger proportion of in-patients among ECT patients (77%) than rTMS patients (8%). In-patients are likely to have more severe symptoms than out-patients. Likely symptoms demanding in-patient care are high risk of suicide and catatonia. Although rTMS and ECT patients were matched for severity of depression as indicated by Clinical Global Impressions Severity scale (CGI-S) score, there may have been aspects of symptom severity that did differ between groups but were not captured by this measure and therefore not taken into account in the analysis. In the ECT group, severity of depression was evaluated by a mixture of clinician-rated MADRS (8%) and patient self-rated MADRS-S (92%), whereas in the rTMS group the severity of depression was evaluated by clinician-rated MADRS only and a single clinician evaluated all cases. The MADRS-S score does not correlate exactly with MADRS score [40], introducing variability between groups. In addition, mild side effects and memory disturbances were not evaluated and recorded equally among cases and controls. In many cases mild side effects may not have been reported to the Q-ECT, hindering a comparison between mild side effects in rTMS and ECT groups. There is always a certain risk for bias in a retrospective case-control study like this, but the results may nevertheless be of scientific and clinical relevance. A randomized controlled study with a similar research question in focus is therefore called upon.

Clinical implications

This study does not provide sufficient evidence of the efficacy of rTMS to warrant any recommendation. Although rTMS seems inferior to ECT in rates of response and remission also in other studies [23,41], the treatment results are interesting. Swedish patients not suited for or willing to partake in ECT for TRD may be helped by rTMS. Advantages of rTMS include no need for anesthesia or post-anesthesia monitoring, which can be limited by the hospital’s availability of anesthesiologist care. More and larger studies comparing ECT and rTMS are needed. Further studies are also required to improve the treatment techniques. Such studies could be facilitated by the inclusion of rTMS in the Q-ECT in 2018.

Conclusions

rTMS may have lower response and remission rates than ECT, although the sample size of this study was too small to make firm conclusions. The side effect profile of rTMS differs from that of ECT, with high rates of mild side effects such as scalp discomfort and headache in the rTMS group, but memory disturbances were only reported in the ECT group. More studies are required to find the optimal place for rTMS within the Swedish health care system.

Acknowledgments

The authors thank Ole Brus, statistician Region Örebro county for help with the matching procedure.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This study was funded by Region Örebro county.