Does electroconvulsive therapy cause brain damage: An update

Abstract

Electroconvulsive therapy (ECT) is an effective modality of treatment for a variety of psychiatric disorders. However, it has always been accused of being a coercive, unethical, and dangerous modality of treatment. The dangerousness of ECT has been mainly attributed to its claimed ability to cause brain damage. This narrative review aims to provide an update of the evidence with regard to whether the practice of ECT is associated with damage to the brain. An accepted definition of brain damage remains elusive. There are also ethical and technical problems in designing studies that look at this question specifically. Thus, even though there are newer technological tools and innovations, any review attempting to answer this question would have to take recourse to indirect methods. These include structural, functional, and metabolic neuroimaging; body fluid biochemical marker studies; and follow-up studies of cognitive impairment and incidence of dementia in people who have received ECT among others. The review of literature and present evidence suggests that ECT has a demonstrable impact on the structure and function of the brain. However, there is a lack of evidence at present to suggest that ECT causes brain damage.

INTRODUCTION

Electroconvulsive therapy (ECT) as a modality of treatment for psychiatric disorders has existed at least since 1938.[1] ECT is an effective modality of treatment for various psychiatric disorders. However, from the very beginning, the practice of ECT has also faced resistance from various groups who claim that it is coercive and harmful.[2] While the ethical aspects of the practice of ECT have been dealt with elsewhere, the question of harmfulness or brain damage consequent upon the passage of electric current needs to be examined afresh in light of technological advances and new knowledge.[3]

The question whether ECT causes brain damage was reviewed in a holistic fashion by Devanand et al. in the mid-1990s.[4,5] The authors had attempted to answer this question by reviewing the effect of ECT on the brain in various areas – cognitive side effects, structural neuroimaging studies, neuropathologic studies of patients who had received ECT, autopsy studies of epileptic patients, and finally animal ECS studies. The authors had concluded that ECT does not produce brain damage.

This narrative review aims to update the evidence with regard to whether ECT causes brain damage by reviewing relevant literature from 1994 to the present time.

FRAMING THE QUESTION

The Oxford Dictionary defines damage as physical harm that impairs the value, usefulness, or normal function of something.[6] Among medical dictionaries, the Peter Collins Dictionary defines damage as harm done to things (noun) or to harm something (verb).[7] Brain damage is defined by the British Medical Association Medical Dictionary as degeneration or death of nerve cells and tracts within the brain that may be localized to a particular area of the brain or diffuse.[8] Going by such a definition, brain damage in the context of ECT should refer to death or degeneration of brain tissue, which results in the impairment of functioning of the brain. The importance of precisely defining brain damage shall become evident subsequently in this review.

There are now many more tools available to investigate the structure and function of brain in health and illness. However, there are obvious ethical issues in designing human studies that are designed to answer this specific question. Therefore, one must necessarily take recourse to indirect evidences available through studies that have been designed to answer other research questions. These studies have employed the following methods:

• Structural neuroimaging studies

• Functional neuroimaging studies

• Metabolic neuroimaging studies

• Body fluid biochemical marker studies

• Cognitive impairment studies.

While the early studies tended to focus more on establishing the safety of ECT and finding out whether ECT causes gross microscopic brain damage, the later studies especially since the advent of advanced neuroimaging techniques have been focusing more on a mechanistic understanding of ECT. Hence, the primary objective of the later neuroimaging studies has been to look for structural and functional brain changes which might explain how ECT acts rather than evidence of gross structural damage per se; however, put together, all these studies would enable us to answer our titular question to some satisfaction. Tables 1 and 2 provide an overview of the evidence base in this area.

Table 1.

Neuroimaging studies in electroconvulsive therapy

Authors	Sample	Method	Result	Methodological issues
				Image quality	Medication status	Control group	Duration of follow-up
Mander et al., 1987[9]	14 patients - depressive illness	MRI before and after ECT - 6 h post to 25 h post ECT	T1 relaxation time increased immediately after ECT - but no long-term increase BBB breakdown - water influx into brain	0.08 Tesla MRI	Not mentioned	5 healthy controls	Last scan - 24 h post ECT
Coffey et al., 1988[13]	9 patients - depressive illness	MRI before and ECT - baseline and 2-3 days after full ECT course	No difference between baseline and follow-up MRI with respect to cortical atrophy, ventricle size, or white matter hyperintensity No structural changes	1.5 Tesla	No psychotropic medication during ECT course	No control group	Last scan 2-3 days after last ECT
Scott et al., 1990[11]	20 patients - depressive illness	Baseline MRI and post-ECT MRI - at 25 min - 2/4/6 and 24 h post-ECT	T1 relaxation time increased immediately post-ECT - but returned to baseline within 24 h	0.08 Tesla	Not mentioned	5 healthy volunteers	Last scan at 24 h after ECT
Pande et al., 1990[12]	7 patients - depressive illness	Baseline MRI and post-ECT MRI - 1 week after ECT course	No new brain changes were found at the post-ECT scanning	0.35 Tesla and 1.5 Tesla MRI	Psychotropic medication discontinued 7 days prior	No control group	Last scan at 1 week post-ECT
Coffey et al., 1991[10]	35 patients - MDD and bipolar depression	Baseline MRI and post-ECT MRI - 2 days after and 6 months after ECT	No significant temporal changes in the total volumes of the lateral ventricles, third ventricles, frontal lobes, temporal lobes, or amygdala-hippocampal complex	1.5 Tesla	In 26 out of 35 patients, psychotropic medications discontinued	No control group	1st - 2 days after last ECT 2nd - 6 months after last ECT
Diehl et al., 1994[14]	6 patients - depressive illness	Baseline MRI and post-ECT MRI - 2 h and 30 h post-ECT	Significant post-ECT T2 relaxation time increase in the right and left thalamus and nonsignificant increase in temporal lobes T2 increase can be attributed to BBB breakdown - leading to edema	1.5 Tesla	Psychotropic medication status not mentioned	No control group	Last imaging done 30 h after 2nd ECT
Kunigiri et al., 2007[15]	15 patients - depressive illness	Baseline MRI and post-ECT - 2 h after 2nd ECT	No significant increase in T2 relaxation time in MRI 2 h post-ECT	1.5 Tesla	Psychotropic medication discontinued	No control group	Last imaging 2 h after 2nd ECT
Szabo et al., 2007[17]	10 patients - MDD, RDD, schizoaffective in depression	Diffusion weighted MRI after ECT - 30 min to 15 h post-ECT MRI	No obvious brain tissue changes seen immediately after ECT - even with DWI	1.5 Tesla	Psychotropic medication status not mentioned	Six healthy elderly controls	Only one MRI imaging done - 2-15 h after ECT
Nordanskog et al., 2010[18]	12 patients - MDD or bipolar depression	3 Tesla MRI 1 week prior to ECT and 1 week post-ECT	Significant increase in hippocampal volume post-ECT - both on left side and right sides and combined	3 Tesla	Patients continued on antidepressants	No control group	Last imaging 1-week post-ECT
Nordanskog et al., 2014[19]	12 patients - unipolar and bipolar depression	MRI 1 week prior to ECT - A1, 1 week after ECT - A2, 6 months post-ECT - A3 and 12 months post-ECT - A4	Significant increase in hippocampal volume at A2 which returned to baseline after 6 months. No clinical correlation between reduction in symptoms and increase in hippocampal volume	3 Tesla	Patients continued on antidepressants	No control group	MRI scans - 1 week, 6 months and 12 months post-ECT
Tendolkar et al., 2013[20]	15 patients - treatment-resistant unipolar depression	MRI 1 week prior to ECT and 1 week post-ECT	Significant increase in the volume of hippocampus and amygdala post-ECT. Increase in volume not significantly correlated with treatment response	1.5 Tesla	Psychotropic medication stopped 1 week prior to ECT	No control group	Last MRI 1 week after last ECT
Dukart et al., 2014[21]	3 groups - 19 with MDD, 15 with BPD 10 patients - 5 MDD + 5 BPD received ECT versus 24 patients who did not receive ECT	Patients assessed before ECT - TP1 and at 3 months post and 6 months post-ECT - TP2 and TP3.	Right unilateral ECT associated with GMV increase in the right side - hippocampus, amygdala, and anterior temporal pole and subgenual cortex ECT associated with decrease in GMV in the right prefrontal cortex. Hippocampal GMV increase and subgenual GMV increase significantly correlated with symptomatic improvement No correlation between ECT dosage and GMV changes or clinical response	1.5 Tesla	Psychotropic medication continued	21 healthy controls	Last MRI 6 months after the last ECT\comment: unilateral ECT - 2.5 times ST administered - not an adequate ECT dosage as per RCPsych Guidelines
Abbott et al., 2014[22]	19 patients - MDD	Patients scanned twice - within 2 days of starting ECT and 5 days after completion of ECT	Right and left hippocampal volume significantly reduced in pre-ECT patients compared to controls Significant increase in hippocampal volume and hippocampal connectivity in ECT responders Right and left hippocampal volumes and connectivity not significantly different between post-ECT responders and healthy controls Increase in hippocampal connectivity correlated with symptom improvement but not so for increase in hippocampal volume	3 Tesla MRI - structural and functional imaging done	Psychotropic medication continued	Twelve healthy controls included - scanned once	Last imaging done 5 days after completion of ECT course
Lyden et al., 2014[23]	20 patients with RDD - presently severe depression	Patients imaged thrice - pre-ECT, 48 h post- first ECT, and after completion of course of ECT	Significant increase in FA and decrease in RD were shown in the anterior cingulum, the forceps minor, and left SLF between pre- and post-ECT groups - suggesting increased fiber integrity in the dorsal fronto-limbic pathway Longitudinal white matter changes were associated with improved clinical response	3 Tesla MRI	Psychotropic medication discontinued	28 healthy controls included - imaged once	Within 1 week of completion of ECT course
Bouckaert et al., 2016[24]	28 patients - MDD, age>55	MRI done - 1 week pre-ECT and 1 week post-ECT	Significant GMV changes only on the side of stimulation in the right unilateral ECT along with a lack of correlation between GMV change and clinical improvement	3 Tesla MRI	Psychotropic medication continued in some patients	No control group	1 week after last ECT in the ECT course
Ota et al., 2015[25]	15 patients with MDD	MRI done roughly 1 week pre-ECT and 1 week post-ECT	Significant increase in volume in bilateral inferior temporal cortices, hippocampi and para hippocampi, right subgenual anterior cingulate gyrus, and right anterior cingulate gyrus post-ECT in total patient group ROI analysis of hippocampal - Amygdala complex: significant increase in volume in the right hippocampus and right amygdala and trend level increase in left amygdala	1.5 Tesla MRI	Psychotropic medication continued in all patients	No control group	1 week after the last ECT
Zeng et al., 2015[26]	24 patients with first-episode MDD	DTI study - patients scanned 1 day before ECT and 1 week after 8 ECTs	Significant reorganization in eight anatomical connections/connectomes - Limbic structure, frontal and temporal lobes - pre- and post-ECT Connection changes between amygdala and para-hippocampus correlated with depressive symptom reduction	3 Tesla MRI	Psychotropic medication continued	No control group	Last MRI 1 week after the last ECT
van Eijndhoven et al., 2016[27]	23 treatment resistant MDD patients - 19 completed study	Patients imaged 1 week before starting ECT and 1 week after completion of ECT - to assess cortical thickness	Large bilateral clusters of increased cortical thickness after ECT treatment - involving the temporal pole, middle and superior temporal cortex, insula, and inferior temporal cortex post-ECT	1.5 Tesla MRI	Psychotropic medication stopped 1 week before beginning ECT	22 healthy controls - 16 completed study	Last MRI 1 week after completion last ECT
Bouckaert et al., 2016[28]	Cohort of 66 MDD patients - 23 remained at the end of 6 months	Imaging done at 1 week before ECT - T0, after 6th ECT - T1, 1 week (T2), 4 weeks (T3) and 6 months (T4) post-ECT Serum BDNF - sBDNF measured at T0, T1, T2, T3, T4 MADRS at T0, T1, T2, T3, T4	ECT induces a significant increase in hippocampal volume - but this does not show correlation with improvement in depressive symptoms. The hippocampal volume increase appears to be transient - returning to the baseline after 6 months - though not accompanied by relapse of depression. Hippcampal volume increase in unrelated to the sBDNF levels. Further, the improvement in depressive symptoms showed no correlation to the sBDNF levels	3 Tesla MRI	Psychotropic medication continued	No control group	Last MRI done at 6 months after ECT (but of the original 66 only 23 remained at 6 months)
Depping et al., 2017[29]	12 ECT-naïve patients with treatment-resistant MDD and 16 healthy controls	Patients imaged 5 days prior to ECT and 8 days post-ECT. Right unilateral ECT administered	At baseline, MDD patients showed increased GMV of right cerebellar area VIIIa and left cerebellar area VIIb compared to healthy controls Post-ECT patients exhibited increased GMV in left cerebellar area VIIa crus I Cerebellar volume increase following ECT was associated with HAMD score reduction	3 Tesla MRI	Psychotropic medication were continued	16 healthy controls	Last MRI 8 days after completion of course of ECT
Joshi et al., 2016[30]	43 patients in severe depression - MDD (n - 35) and bipolar disorder (n - 7)	Patients assessed at three time points - 24 h before ECT - T1, 24 h after 3rd ECT - T2, within 1 week of completing ECT - T3	Hippocampal and amygdalar volume were significantly lesser in pre-ECT patients compared to healthy controls and also that there was a significant increase in the volume of bilateral hippocampus and amygdala post-ECT Significant correlation between increase in hippocampal and amygdalar volume and reduction in HAMD	3 Tesla MRI	Psychotropic medication discontinued	32 healthy controls imaged twice	Last MRI within 1 week of last ECT
Pirnia et al., 2016[16]	Patients with MDD or BPAD currently severe depression - 41 enrolled - 29 completed	Patients imaged at three time points - T1 - baseline before ECT, T2 - 48 h after 2nd ECT, T3 - within 1 week of completing ECT	Pair-wise comparisons at T1 and T3 showed significant changes in the bilateral ACC, right para - hippocampal gyrus, superior temporal gyrus, and temporal pole ROI analysis showed significant increase in cortical thickness post-ECT in ACC, para hippocampal, entorhinal, superior temporal, inferior temporal, and fusiform cortex Significant correlation between increase in cortical thickness in the fusiform, superior, inferior temporal cortex and reduction in clinical symptoms	3 Tesla MRI	Psychotropic medication discontinued	29 healthy controls	Last MRI within 1 week of last ECT
Wade et al., 2016[31]	53 patients in severe depression - unipolar or bipolar selected - 34 completed all assessments	Patients assessed at three time points - T1 before ECT, T2 - 48 h after 2nd ECT and T3 - 1 week after last ECT	At baseline - nucleus acumbens and pallidum volume significantly reduced in patients. Volume increase of putamen during course of ECT. Differential response of nucleus acumbens and caudate nucleus to ECT in responders and nonresponders. Patient’s responsiveness to ECT predicted with 89% accuracy using machine learning techniques	3 Tesla MRI	Psychotropic medication discontinued prior to starting ECT	33 healthy controls - controls assessed at two time points: 4 - 6 weeks apart	Last MRI 1 week after last ECT
Wolf et al., 2016[32]	21 ECT-naïve patients - 12 with MDD and 9 with schizophrenia	Patients imaged twice - within 5 days prior to the first ECT session and 6-8 days after last ECT	ECT leads to increase in structural network strength in MPFC and MTL networks in MDD - but no correlation with clinical improvement In schizophrenia patients with depression - significant difference post-ECT in two networks - DLPFC and MTL network with significant correlation between clinical improvement and the increase in strength of the DLPFC network	3 Tesla MRI	Psychotropic medication continued	21 healthy controls	Last MRI 8 days after last ECT
Qiu et al., 2016[14]	12 patients - Unipolar depression and 15 matched controls	Imaging 1 day pre- and post-ECT.	Comparison between pre- and post-ECT data showed increase in GMV in bilateral hippocampus and amygdala Analysis of fMRI data indicated baseline functional differences between healthy controls and MDD patients with functional changes occurring in MDD patients post-ECT - though not returning to full normalcy	3 Tesla MRI	Psychotropic medication discontinued	15 healthy controls	Last MRI 1 day after the 8th ECT in a series of 12 ECTs

MDD – Major depressive disorder; RDD – Recurrent depressive disorder; BPD – Borderline personality disorder; ECTs – Electroconvulsive therapies; BPAD – Bipolar affective disorder; MRI – Magnetic resonance imaging; BDNF – Brain-derived neurotrophic factor; sBDNF – Serum BDNF; DTI – Diffusion tensor imaging; fMRI – Functional MRI; BBB – Blood–brain barrier; DWI – Diffusion-weighted imaging; GMV – Gray matter volume; FA – Fractional anisotropy; RD – Radial diffusivity; SLF – Superior longitudinal fasciculus; HAMD – Hamilton Depression Rating Scale; ACC – Anterior cingulate cortex; MPFC – Medial prefrontal cortex; MTL – Medial temporal lobe; DLPFC – Dorsolateral prefrontal cortex; ST – Seizure Threshold; ROI – Region of Interest

Table 2.

Neurochemical studies in electroconvulsive therapy

Authors	Sample	Method	Result	Limitation
				Image quality	Medication status	Control group	Follow-up
Ende et al., 2000[33]	17 patients - depressive illness	Baseline MRSI - post-ECT: After completion of at least five ECTs	No change in hippocampal NAA signals - reduced NAA - marker of neuronal death - hence, no neuronal death Significant increase in hippocampal choline signals - marker of membrane turnover - indicates mossy fiber sprouting in hippocampus	1.5 Tesla MRSI	Psychotropic medication discontinued	24 healthy controls	Last MRI 30 h-10 days after 5th or later ECT
Michael et al., 2003[34]	28 patients fulfilling DSM IV TR criteria for MDD: 13 unipolar depression and 15 bipolar depression	Left amygdalar region studied using 1.5 T steam MRSI Imaging done before and after right unilateral ECT at 2.5 ST (one case - bilateral ECT)	Before ECT, the concentration of NAA, Cho, and Cr was not statistically different between patients and controls Patients with unipolar depression showed a significantly lower concentration of Glx Successful ECT was accompanied by significant increase in NAA and Glx levels Cho and Cr increased without reaching statistical significance	1.5 Tesla MRSI	Psychotropic medication discontinued	28 healthy controls	Exact duration of follow-up not mentioned - no long-term follow-up
Michael et al., 2003[35]	12 patients with severe depression	Baseline MRSI and post-ECT MRSI (average 30 h after ECT)	Depressed patients had significantly reduced Glutamate/Glutamine - Glx levels in the DLPFC ECT responders - Marked increase in Glx levels in DLPFC ECT nonresponders - No significant increase in Glx levels in DLPFC No significant difference in NAA levels between patients and controls	1.5 Tesla MRSI	Psychotropic medication discontinued	12 age- and gender-matched healthy controls	Last MRI approximately 30 h after ECT - No long-term follow-up
Pfleiderer et al., 2003[36]	17 patients - severe recurrent unipolar depression	Medication washout period of 3-8 days 1.5 T MRSI Baseline MRSI and post-ECT MRSI (24-48 h) Unilateral ECT - 2.5 ST: 15 patients bilateral ECT - 2 ST: 2 patients	Depressed patients had significantly reduced Glx levels in the left anterior cingulate gyrus ECT responders had significant increase in Glx levels in the left anterior cingulate gyrus. ECT nonresponders - no significant increase in left anterior cingulate gyrus No significant difference in Glx levels between ECT responders and controls post-ECT - levels equalized No significant difference in levels of NAA, Cho, and Cr between patients and controls pre-ECT. No significant elevation of levels of NAA, Cho, or Cr post-ECT	1.5 Tesla MRSI	Psychotropic medication discontinued	17 age- and gender-matched healthy controls	Second MRI 24-48 h after last ECT - No long-term follow-up
Merkl et al., 2011[37]	27 patients with MDD	Right unilateral ultra-brief pulse ECT at 4 ST, 7 ST, and 10 ST - Randomized two patients received bilateral ECT 3 Tesla MRSI pre- and post-ECT		3 Tesla MRSI	Psychotropic medication continued	27 age- and gender-matched healthy controls	Second MRI after nine ECTs - No long-term follow-up
			Significant reduction in NAA and glutamate levels in the left anterior cingulum in patients
			No significant difference in Cho or Cr levels between patients and controls - both in DLPFC and anterior cingulum Significant reduction in NAA in left DLPFC post-ECT in responders Significant increase in NAA levels in the left anterior cingulum post-ECT in responders High glutamate level in the left anterior cingulum at baseline predicted greater improvement with ECT No significant correlation between glutamate levels and stimulus dosage of 4 ST, 7 ST, and 10 ST
Jorgensen et al., 2016[38]	19 patients with unipolar or bipolar severe depression	Patients assessed at three time points - T1 - before ECT: MRI, blood sample, T2 - 1 week after ECT: MRI, T3 - 4 weeks after ECT: MRI Bilateral bitemporal ECT at 1.5 ST switched to right unilateral ECT in case of severe cognitive impairment	Highly significant increase in brain volume for bilateral hippocampus and amygdala between T1 and T2 and also to a lesser extent between T1 and T3. Significant reduction in brain volume for DLPC between T1 and T3. Significant reduction volume of DLFPC post-ECT. No significant correlation between increase in volume and clinical improvement Significant decrease in FA for hippocampus between T1 and T2 and significant increase in FA for hypothalamus between T1 and T2 Also, significant reduction in MD for both hippocampus and hypothalamus between T1 and T2 No correlation between changes in FA or MD and clinical response MRSI of the left and right hippocampus showed no significant change in the concentration of NAA, Cho, and Cr between T1, T2 and T3 - however, nonsignificant increase in Cr+PCr values in the left hippocampus post-ECT Nonsignificant increase in serum BDNF levels between T1 and T2, which is lost at T3. No correlation between elevation of serum BDNF levels and clinical improvement. No correlation between serum BDNF levels and increase in regional brain volumes	3 Tesla MRSI	Psychotropic medication continued	No healthy control group	Last MRI 4 weeks after the last ECT
Njau et al., 2017[39]	50 patients with severe depression - unipolar or bipolar with a history of at least two prior episodes and at least two antidepressant trails	Patients assessed three times: T1 - 24 h prior to ECT, T2 - After 2nd ECT and T3 - 1 week after last ECT Controls scanned twice - Baseline and 4 weeks later Bilateral ECT at 1.5 ST and right unilateral ECT at 5 ST	Regions studied: dACC and sgACC and bilateral hippocampus Prior to ECT: Significantly reduced NAA and elevated Glx in the left hippocampus in patients. Significantly reduced Glx levels in sgACC in patients compared to controls Longitudinal effect of ECT: Right hippocampal NAA reduction and left hippocampal Glx reduction. dACC showed decreased NAA and increased Cr. sgACC showed increased Glx and Cr Correlation with clinical improvement - for decrease in Glx in left hippocampus and increase in Glx in sgACC Predictor of response: Higher baseline levels of NAA in dACC predicts greater response	3 Tesla MRSI	Psychotropic medication discontinued	33 age- and gender matched healthy controls	Last MRI 1 week after last ECT
Cano et al., 2017[40]	12 patients with treatment-resistant MDD	Patients were scanned four times - 24 h prior to ECT, 24 h after first ECT, 24 h after 9th ECT, and 2 weeks after completion of ECT course Controls were scanned twice - 5 weeks apart bilateral fronto-temporal ECT - stimulus dose by half-age method	Pre-ECT: No structural or metabolite concentration difference between the patient and control groups Longitudinal change: Significant increase in bilateral MTL which includes: hippocampus, parahippocampus, and amygdala and rPgACC Hippocampal NAA/Cr ratio significantly reduced after the 9th ECT and Glx/Cr ratio increased at a trend level at the same point. Cho/Cr ratio did not significantly vary between the time points Volume and metabolite correlation: Significant correlation between left hippocampal volume change and bilateral hippocampal NAA/Cr ratio change and a trend level correlation between left hippocampal volume change and bilateral hippocampal Glx/Cr ratio change - nonsignificant after Bonferroni correction for multiple comparisons Clinical correlation: Significant correlation between left MTL volume change and clinical improvement. No significant correlation between metabolite concentration change and clinical improvement	3 Tesla MRSI	Psychotropic medication continued	10 age- and gender-matched healthy controls	Last MRI 2 weeks after the last ECT

MDD – Major depressive disorder; ECTs – Electroconvulsive therapys; MRSI – Magnetic resonance spectroscopic imaging; NAA – N-acetyl aspartate; Cho – Choline-containing compounds; GLX – Glutamine and glutamate together; Cr – Creatinine; DLPFC – Dorsolateral prefrontal cortex; FA – Fractional anisotropy; MD – Mean diffusivity; BDNF – Brain-derived neurotrophic factor; ACC – Anterior cingulate cortex; dACC – Dorsal ACC; sgACC – Subgenual ACC; GMV – Gray matter volume; rPgACC – Right perigenual ACC; MTL – Medial temporal lobe; DSM IV TR – Diagnostic and Statistical Manual IV – Text Revision; ST – Seizure Threshold

STRUCTURAL AND FUNCTIONAL NEUROIMAGING STUDIES

Devanand et al. reviewed 16 structural neuroimaging studies on the effect of ECT on the brain.[4] Of these, two were pneumoencephalography studies, nine were computed tomography (CT) scan studies, and five were magnetic resonance imaging (MRI) studies. However, most of these studies were retrospective in design, with neuroimaging being done in patients who had received ECT in the past. In the absence of baseline neuroimaging, it would be very difficult to attribute any structural brain changes to ECT. In addition, pneumoencephalography, CT scan, and even early 0.3 T MRI provided images with much lower spatial resolution than what is available today. The authors concluded that there was no evidence to show that ECT caused any structural damage to the brain.[4] Since then, at least twenty more MRI-based structural neuroimaging studies have studied the effect of ECT on the brain. The earliest MRI studies in the early 1990s focused on detecting structural damage following ECT. All of these studies were prospective in design, with the first MRI scan done at baseline and a second MRI scan performed post ECT.[9,11,12,13,41] While most of the studies imaged the patient once around 24 h after receiving ECT, some studies performed multiple post ECT neuroimaging in the first 24 h after ECT to better capture the acute changes. A single study by Coffey et al. followed up the patients for a duration of 6 months and repeated neuroimaging again at 6 months in order to capture any long-term changes following ECT.[10]

The most important conclusion which emerged from this early series of studies was that there was no evidence of cortical atrophy, change in ventricle size, or increase in white matter hyperintensities.[4] The next major conclusion was that there appeared to be an increase in the T1 and T2 relaxation time immediately following ECT, which returned to normal within 24 h. This supported the theory that immediately following ECT, there appears to be a temporary breakdown of the blood–brain barrier, leading to water influx into the brain tissue.[11] The last significant observation by Coffey et al. in 1991 was that there was no significant temporal changes in the total volumes of the frontal lobes, temporal lobes, or amygdala–hippocampal complex.[10] This was, however, something which would later be refuted by high-resolution MRI studies. Nonetheless, one inescapable conclusion of these early studies was that there was no evidence of any gross structural brain changes following administration of ECT. Much later in 2007, Szabo et al. used diffusion-weighted MRI to image patients in the immediate post ECT period and failed to observe any obvious brain tissue changes following ECT.[17]

The next major breakthrough came in 2010 when Nordanskog et al. demonstrated that there was a significant increase in the volume of the hippocampus bilaterally following a course of ECT in a cohort of patients with depressive illness.[18] This contradicted the earlier observations by Coffey et al. that there was no volume increase in any part of the brain following ECT.[10] This was quite an exciting finding and was followed by several similar studies. However, the perspective of these studies was quite different from the early studies. In contrast to the early studies looking for the evidence of ECT-related brain damage, the newer studies were focused more on elucidating the mechanism of action of ECT. Further on in 2014, Nordanskog et al. in a follow-up study showed that though there was a significant increase in the volume of the hippocampus 1 week after a course of ECT, the hippocampal volume returned to the baseline after 6 months.[19] Two other studies in 2013 showed that in addition to the hippocampus, the amygdala also showed significant volume increase following ECT.[20,21] A series of structural neuroimaging studies after that have expanded on these findings and as of now, gray matter volume increase following ECT has been demonstrated in the hippocampus, amygdala, anterior temporal pole, subgenual cortex,[21] right caudate nucleus, and the whole of the medial temporal lobe (MTL) consisting of the hippocampus, amygdala, insula, and the posterosuperior temporal cortex,[24] para hippocampi, right subgenual anterior cingulate gyrus, and right anterior cingulate gyrus,[25] left cerebellar area VIIa crus I,[29] putamen, caudate nucleus, and nucleus acumbens[31] and clusters of increased cortical thickness involving the temporal pole, middle and superior temporal cortex, insula, and inferior temporal cortex.[27] However, the most consistently reported and replicated finding has been the bilateral increase in the volume of the hippocampus and amygdala. In light of these findings, it has been tentatively suggested that ECT acts by inducing neuronal regeneration in the hippocampus – amygdala complex.[42,43] However, there are certain inconsistencies to this hypothesis. Till date, only one study – Nordanskog et al., 2014 – has followed study patients for a long term – 6 months in their case. And significantly, the authors found out that after increasing immediately following ECT, the hippocampal volume returns back to baseline by 6 months.[19] This, however, was not associated with the relapse of depressive symptoms. Another area of significant confusion has been the correlation of hippocampal volume increase with improvement of depressive symptoms. Though almost all studies demonstrate a significant increase in hippocampal volume following ECT, a majority of studies failed to demonstrate a correlation between symptom improvement and hippocampal volume increase.[19,20,22,24,28] However, a significant minority of volumetric studies have demonstrated correlation between increase in hippocampal and/or amygdala volume and improvement of symptoms.[21,25,30]

Another set of studies have used diffusion tensor imaging, functional MRI (fMRI), anatomical connectome, and structural network analysis to study the effect of ECT on the brain. The first of these studies by Abbott et al. in 2014 demonstrated that on fMRI, the connectivity between right and left hippocampus was significantly reduced in patients with severe depression. It was also shown that the connectivity was normalized following ECT, and symptom improvement was correlated with an increase in connectivity.[22] In a first of its kind DTI study, Lyden et al. in 2014 demonstrated that fractional anisotropy which is a measure of white matter tract or fiber density is increased post ECT in patients with severe depression in the anterior cingulum, forceps minor, and the dorsal aspect of the left superior longitudinal fasciculus. The authors suggested that ECT acts to normalize major depressive disorder-related abnormalities in the structural connectivity of the dorsal fronto-limbic pathways.[23] Another DTI study in 2015 constructed large-scale anatomical networks of the human brain – connectomes, based on white matter fiber tractography. The authors found significant reorganization in the anatomical connections involving the limbic structure, temporal lobe, and frontal lobe. It was also found that connection changes between amygdala and para hippocampus correlated with reduction in depressive symptoms.[26] In 2016, Wolf et al. used a source-based morphometry approach to study the structural networks in patients with depression and schizophrenia and the effect of ECT on the same. It was found that the medial prefrontal cortex/anterior cingulate cortex (ACC/MPFC) network, MTL network, bilateral thalamus, and left cerebellar regions/precuneus exhibited significant difference between healthy controls and the patient population. It was also demonstrated that administration of ECT leads to significant increase in the network strength of the ACC/MPFC network and the MTL network though the increase in network strength and symptom amelioration were not correlated.[32]

Building on these studies, a recently published meta-analysis has attempted a quantitative synthesis of brain volume changes – focusing on hippocampal volume increase following ECT in patients with major depressive disorder and bipolar disorder. The authors initially selected 32 original articles from which six articles met the criteria for quantitative synthesis. The results showed significant increase in the volume of the right and left hippocampus following ECT. For the rest of the brain regions, the heterogeneity in protocols and imaging techniques did not permit a quantitative analysis, and the authors have resorted to a narrative review similar to the present one with similar conclusions.[44] Focusing exclusively on hippocampal volume change in ECT, Oltedal et al. in 2018 conducted a mega-analysis of 281 patients with major depressive disorder treated with ECT enrolled at ten different global sites of the Global ECT-MRI Research Collaboration.[45] Similar to previous studies, there was a significant increase in hippocampal volume bilaterally with a dose–response relationship with the number of ECTs administered. Furthermore, bilateral (B/L) ECT was associated with an equal increase in volume in both right and left hippocampus, whereas right unilateral ECT was associated with greater volume increase in the right hippocampus. Finally, contrary to expectation, clinical improvement was found to be negatively correlated with hippocampal volume.

Thus, a review of the current evidence amply demonstrates that from looking for ECT-related brain damage – and finding none, we have now moved ahead to looking for a mechanistic understanding of the effect of ECT. In this regard, it has been found that ECT does induce structural changes in the brain – a fact which has been seized upon by some to claim that ECT causes brain damage.[46] Such statements should, however, be weighed against the definition of damage as understood by the scientific medical community and patient population. Neuroanatomical changes associated with effective ECT can be better described as ECT-induced brain neuroplasticity or ECT-induced brain neuromodulation rather than ECT-induced brain damage.

METABOLIC NEUROIMAGING STUDIES: MAGNETIC RESONANCE SPECTROSCOPIC IMAGING

Magnetic resonance spectroscopic imaging (MRSI) uses a phase-encoding procedure to map the spatial distribution of magnetic resonance (MR) signals of different molecules. The crucial difference, however, is that while MRI maps the MR signals of water molecules, MRSI maps the MR signals generated by different metabolites – such as N-acetyl aspartate (NAA) and choline-containing compounds. However, the concentration of these metabolites is at least 10,000 times lower than water molecules and hence the signal strength generated would also be correspondingly lower. However, MRSI offers us the unique advantage of studying in vivo the change in the concentration of brain metabolites, which has been of great significance in fields such as psychiatry, neurology, and basic neuroscience research.[47]

MRSI studies on ECT in patients with depression have focused largely on four metabolites in the human brain – NAA, choline-containing compounds (Cho) which include majorly cell membrane compounds such as glycerophosphocholine, phosphocholine and a miniscule contribution from acetylcholine, creatinine (Cr) and glutamine and glutamate together (Glx). NAA is located exclusively in the neurons, and is suggested to be a marker of neuronal viability and functionality.[48] Choline-containing compounds (Cho) mainly include the membrane compounds, and an increase in Cho would be suggestive of increased membrane turnover. Cr serves as a marker of cellular energy metabolism, and its levels are usually expected to remain stable. The regions which have been most widely studied in MRSI studies include the bilateral hippocampus and amygdala, dorsolateral prefrontal cortex (DLPFC), and ACC.

Till date, five MRSI studies have measured NAA concentration in the hippocampus before and after ECT. Of these, three studies showed that there is no significant change in the NAA concentration in the hippocampus following ECT.[33,38,49] On the other hand, two recent studies have demonstrated a statistically significant reduction in NAA concentration in the hippocampus following ECT.[39,40] The implications of these results are of significant interest to us in answering our titular question. A normal level of NAA following ECT could signify that there is no significant neuronal death or damage following ECT, while a reduction would signal the opposite. However, a direct comparison between these studies is complicated chiefly due to the different ECT protocols, which has been used in these studies. It must, however, be acknowledged that the three older studies used 1.5 T MRI, whereas the two newer studies used a higher 3 T MRI which offers betters signal-to-noise ratio and hence lesser risk of errors in the measurement of metabolite concentrations. The authors of a study by Njau et al.[39] argue that a change in NAA levels might reflect reversible changes in neural metabolism rather than a permanent change in the number or density of neurons and also that reduced NAA might point to a change in the ratio of mature to immature neurons, which, in fact, might reflect enhanced adult neurogenesis. Thus, the authors warn that to conclude whether a reduction in NAA concentration is beneficial or harmful would take a simultaneous measurement of cognitive functioning, which was lacking in their study. In 2017, Cano et al. also demonstrated a significant reduction in NAA/Cr ratio in the hippocampus post ECT. More significantly, the authors also showed a significant increase in Glx levels in the hippocampus following ECT, which was also associated with an increase in hippocampal volume.[40] To explain these three findings, the authors proposed that ECT produces a neuroinflammatory response in the hippocampus – likely mediated by Glx, which has been known to cause inflammation at higher concentrations, thereby accounting for the increase in hippocampal volume with a reduction in NAA concentration. The cause for the volume increase remains unclear – with the authors speculating that it might be due to neuronal swelling or due to angiogenesis. However, the same study and multiple other past studies[21,25,30] have demonstrated that hippocampal volume increase was correlated with clinical improvement following ECT. Thus, we are led to the hypothesis that the same mechanism which drives clinical improvement with ECT is also responsible for the cognitive impairment following ECT. Whether this is a purely neuroinflammatory response or a neuroplastic response or a neuroinflammatory response leading to some form of neuroplasticity is a critical question, which remains to be answered.[40]

Studies which have analyzed NAA concentration change in other brain areas have also produced conflicting results. The ACC is another area which has been studied in some detail utilizing the MRSI technique. In 2003, Pfleiderer et al. demonstrated that there was no significant change in the NAA and Cho levels in the ACC following ECT. This would seem to suggest that there was no neurogenesis or membrane turnover in the ACC post ECT.[36] However, this finding was contested by Merkl et al. in 2011, who demonstrated that NAA levels were significantly reduced in the left ACC in patients with depression and that these levels were significantly elevated following ECT.[37] This again is contested by Njau et al. who showed that NAA levels are significantly reduced following ECT in the left dorsal ACC.[39] A direct comparison of these three studies is complicated by the different ECT and imaging parameters used and hence, no firm conclusion can be made on this point at this stage. In addition to this, one study had demonstrated increased NAA levels in the amygdala following administration of ECT,[34] with a trend level increase in Cho levels, which again is suggestive of neurogenesis and/or neuroplasticity. A review of studies on the DLPFC reveals a similarly confusing picture with one study, each showing no change, reduction, and elevation of concentration of NAA following ECT.[35,37,39] Here, again, a direct comparison of the three studies is made difficult by the heterogeneous imaging and ECT protocols followed by them.

A total of five studies have analyzed the concentration of choline-containing compounds (Cho) in patients undergoing ECT. Conceptually, an increase in Cho signals is indicative of increased membrane turnover, which is postulated to be associated with synaptogenesis, neurogenesis, and maturation of neurons.[31] Of these, two studies measured Cho concentration in the B/L hippocampus, with contrasting results. Ende et al. in 2000 demonstrated a significant elevation in Cho levels in B/L hippocampus after ECT, while Jorgensen et al. in 2015 failed to replicate the same finding.[33,38] Cho levels have also been studied in the amygdala, ACC, and the DLPFC. However, none of these studies showed a significant increase or decrease in Cho levels before and after ECT in the respective brain regions studied. In addition, no significant difference was seen in the pre-ECT Cho levels of patients compared to healthy controls.[34,36,37]

In review, we must admit that MRSI studies are still at a preliminary stage with significant heterogeneity in ECT protocols, patient population, and regions of the brain studied. At this stage, it is difficult to draw any firm conclusions except to acknowledge the fact that the more recent studies – Njau et al., 2017, Cano, 2017, and Jorgensen et al., 2015 – have shown decrease in NAA concentration and no increase in Cho levels[38,39,40] – as opposed to the earlier studies by Ende et al.[33] The view offered by the more recent studies is one of a neuroinflammatory models of action of ECT, probably driving neuroplasticity in the hippocampus. This would offer a mechanistic understanding of both clinical response and the phenomenon of cognitive impairment associated with ECT. However, this conclusion is based on conjecture, and more work needs to be done in this area.

BODY FLUID BIOCHEMICAL MARKER STUDIES

Another line of evidence for analyzing the effect of ECT on the human brain is the study of concentration of neurotrophins in the plasma or serum. Neurotrophins are small protein molecules which mediate neuronal survival and development. The most prominent among these is brain-derived neurotrophic factor (BDNF) which plays an important role in neuronal survival, plasticity, and migration.[50] A neurotrophic theory of mood disorders was suggested which hypothesized that depressive disorders are associated with a decreased expression of BDNF in the limbic structures, resulting in the atrophy of these structures.[51] It was also postulated that antidepressant treatment has a neurotrophic effect which reverses the neuronal cell loss, thereby producing a therapeutic effect. It has been well established that BDNF is decreased in mood disorders.[52] It has also been shown that clinical improvement of depression is associated with increase in BDNF levels.[53] Thus, serum BDNF levels have been tentatively proposed as a biomarker for treatment response in depression. Recent meta-analytic evidence has shown that ECT is associated with significant increase in serum BDNF levels in patients with major depressive disorder.[54] Considering that BDNF is a potent stimulator of neurogenesis, the elevation of serum BDNF levels following ECT lends further credence to the theory that ECT leads to neurogenesis in the hippocampus and other limbic structures, which, in turn, mediates the therapeutic action of ECT.

COGNITIVE IMPAIRMENT STUDIES

Cognitive impairment has always been the single-most important side effect associated with ECT.[55] Concerns regarding long-term cognitive impairment surfaced soon after the introduction of ECT and since then has grown to become one of the most controversial aspects of ECT.[56] Anti-ECT groups have frequently pointed out to cognitive impairment following ECT as evidence of ECT causing brain damage.[56] A meta-analysis by Semkovska and McLoughlin in 2010 is one of the most detailed studies which had attempted to settle this long-standing debate.[57] The authors reviewed 84 studies (2981 participants), which had used a combined total of 22 standardized neuropsychological tests assessing various cognitive functions before and after ECT in patients diagnosed with major depressive disorder. The different cognitive domains reviewed included processing speed, attention/working memory, verbal episodic memory, visual episodic memory, spatial problem-solving, executive functioning, and intellectual ability. The authors concluded that administration of ECT for depression is associated with significant cognitive impairment in the first few days after ECT administration. However, it was also seen that impairment in cognitive functioning resolved within a span of 2 weeks and thereafter, a majority of cognitive domains even showed mild improvement compared to the baseline performance. It was also demonstrated that not a single cognitive domain showed persistence of impairment beyond 15 days after ECT.

Memory impairment following ECT can be analyzed broadly under two conceptual schemes – one that classifies memory impairment as objective memory impairment and subjective memory impairment and the other that classifies it as impairment in anterograde memory versus impairment in retrograde memory. Objective memory can be roughly defined as the ability to retrieve stored information and can be measured by various standardized neuropsychological tests. Subjective memory or meta-memory, on the other hand, refers to the ability to make judgments about one's ability to retrieve stored information.[58] As described previously, it has been conclusively demonstrated that anterograde memory impairment does not persist beyond 2 weeks after ECT.[57] However, one of the major limitations of this meta-analysis was the lack of evidence on retrograde amnesia following ECT. This is particularly unfortunate considering that it is memory impairment – particularly retrograde amnesia which has received the most attention.[59] In addition, reports of catastrophic retrograde amnesia have been repeatedly held up as sensational evidence of the lasting brain damage produced by ECT.[59] Admittedly, studies on retrograde amnesia are fewer and less conclusive than on anterograde amnesia.[60,61] At present, the results are conflicting, with some studies finding some impairment in retrograde memory – particularly autobiographical retrograde memory up to 6 months after ECT.[62,63,64,65] However, more recent studies have failed to support this finding.[66,67] While they do demonstrate an impairment in retrograde memory immediately after ECT, it was seen that this deficit returned to pre-ECT levels within a span of 1–2 months and improved beyond baseline performance at 6 months post ECT.[66] Adding to the confusion are numerous factors which confound the assessment of retrograde amnesia. It has been shown that depressive symptoms can produce significant impairment of retrograde memory.[68,69] It has also been demonstrated that sine-wave ECT produces significantly more impairment of retrograde memory as compared to brief-pulse ECT.[70] However, from the 1990s onward, sine-wave ECT has been completely replaced by brief-pulse ECT, and it is unclear as to the implications of cognitive impairment from the sine-wave era in contemporary ECT practice.

Another area of concern are reports of subjective memory impairment following ECT. One of the pioneers of research into subjective memory impairment were Squire and Chace who published a series of studies in the 1970s demonstrating the adverse effect of bilateral ECT on subjective assessment of memory.[62,63,64,65] However, most of the studies conducted post 1980 – from when sine-wave ECT was replaced by brief-pulse ECT report a general improvement in subjective memory assessments following ECT.[71] In addition, most of the recent studies have failed to find a significant association between measures of subjective and objective memory.[63,66,70,72,73,74] It has also been shown that subjective memory impairment is strongly associated with the severity of depressive symptoms.[75] In light of these facts, the validity and value of measures of subjective memory impairment as a marker of cognitive impairment and brain damage following ECT have been questioned. However, concerns regarding subjective memory impairment and catastrophic retrograde amnesia continue to persist, with significant dissonance between the findings of different research groups and patient self-reports in various media.[57]

Some studies reported the possibility of ECT being associated with the development of subsequent dementia.[76,77] However, a recent large, well-controlled prospective Danish study found that the use of ECT was not associated with elevated incidence of dementia.[78]

CONCLUSION

Our titular question is whether ECT leads to brain damage, where damage indicates destruction or degeneration of nerves or nerve tracts in the brain, which leads to loss of function. This issue was last addressed by Devanand et al. in 1994 since which time our understanding of ECT has grown substantially, helped particularly by the advent of modern-day neuroimaging techniques which we have reviewed in detail. And, what these studies reveal is rather than damaging the brain, ECT has a neuromodulatory effect on the brain. The various lines of evidence – structural neuroimaging studies, functional neuroimaging studies, neurochemical and metabolic studies, and serum BDNF studies all point toward this. These neuromodulatory changes have been localized to the hippocampus, amygdala, and certain other parts of the limbic system. How exactly these changes mediate the improvement of depressive symptoms is a question that remains unanswered. However, there is little by way of evidence from neuroimaging studies which indicates that ECT causes destruction or degeneration of neurons. Though cognitive impairment studies do show that there is objective impairment of certain functions – particularly memory immediately after ECT, these impairments are transient with full recovery within a span of 2 weeks. Perhaps, the single-most important unaddressed concern is retrograde amnesia, which has been shown to persist for up to 2 months post ECT. In this regard, the recent neurometabolic studies have offered a tentative mechanism of action of ECT, producing a transient inflammation in the limbic cortex, which, in turn, drives neurogenesis, thereby exerting a neuromodulatory effect. This hypothesis would explain both the cognitive adverse effects of ECT – due to the transient inflammation – and the long-term improvement in mood – neurogenesis in the hippocampus. Although unproven at present, such a hypothesis would imply that cognitive impairment is tied in with the mechanism of action of ECT and not an indicator of damage to the brain produced by ECT.

The review of literature suggests that ECT does cause at least structural and functional changes in the brain, and these are in all probability related to the effects of the ECT. However, these cannot be construed as brain damage as is usually understood. Due to the relative scarcity of data that directly examines the question of whether ECT causes brain damage, it is not possible to conclusively answer this question. However, in light of enduring ECT survivor accounts, there is a need to design studies that specifically answer this question.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.