Efficacy and adverse effects of ketamine versus electroconvulsive therapy for major depressive disorder: A systematic review and meta-analysis

Abstract

Background:

ECT is considered the fastest and most effective treatment for TRD. Ketamine seems to be an attractive alternative due to its rapid-onset antidepressant effects and impact on suicidal thoughts. This study aimed to compare efficacy and tolerability of ECT and ketamine for different depression outcomes (PROSPERO/CRD42022349220).

Methods:

We searched MEDLINE, Web of Science, Embase, PsycINFO, Google Scholar, Cochrane Library and trial registries, which were the ClinicalTrials.gov and the World Health Organization’s International Clinical Trials Registry Platform, without restrictions on publication date. Selection criteria: randomized controlled trials or cohorts comparing ketamine versus ECT in patients with TRD.

Results:

Eight studies met the inclusion criteria (of 2875 retrieved). Random-effects models comparing ketamine and ECT regarding the following outcomes were conducted: a) reduction of depressive symptoms severity through scales, g = −0.12, p = 0.68; b) response to therapy, RR = 0.89, p = 0.51; c) reported side-effects: dissociative symptoms, RR = 5.41, p = 0.06; nausea, RR = 0.73, p = 0.47; muscle pain, RR = 0.25, p = 0.02 and headache, RR = 0.39, p = 0.08. Influential & subgroup analyses were performed.

Limitations:

Methodological issues with high risk of bias in some of the source material, reduced number of eligible studies with high in-between heterogeneity and small sample sizes.

Conclusion:

Our study showed no evidence to support the superiority of ketamine over ECT for severity of depressive symptoms and response to therapy. Regarding side effects, there was a statistically significant decreased risk of muscle pain in patients treated with ketamine compared to ECT.

1. Background

According to DSM-V, major depressive disorder (MDD) criteria is met by the presence of at least five of the following symptoms occurring independent of physical illness, normal bereavement, alcohol or drugs: abnormal depressed mood; abnormal loss of interest and pleasure; appetite or weight disturbance; sleep disturbance; disturbance in activity (agitation or slowing); abnormal fatigue or loss of energy; abnormal self-reproach or inappropriate guilt; poor concentration or indecisiveness; morbid thoughts of death or suicide (American Psychiatric Association, 2013). At least one symptom must be abnormal depressed mood or loss of interest and pleasure, persisting for most of nearly every day for at least two weeks and significantly impairing function and daily life (Pandarakalam, 2018). It constitutes one of the most prevalent (16 %–20 %) psychiatric disorders worldwide (Sadock et al., 2015). Common therapeutic strategies include pharmacotherapy (which acts through serotonin, dopamine, and norepinephrine systems). However, a reduction in depressive symptoms is observed within several weeks after the start of treatment after conventional antidepressants. Regardless, remission with this therapy remains insufficient after several weeks, and one-third of patients fail to achieve functional recovery despite multimodality treatment interventions (Rong et al., 2018).

There has been considerable debate regarding what constitutes treatment-resistant depression (TRD), and whether medications from more than one class must be trialed prior to meeting criteria for this classification (Voineskos et al., 2020). However, it is generally classified as TRD once two adequate antidepressant trials have been unsuccessful (Rush et al., 2006). Sullivan (1995) included treatment resistance levels ranging from one failed antidepressant trial to a lack of response to electroconvulsive therapy (ECT).

The gold standard treatment for TRD is ECT (Merkl et al., 2009). It acts via a controlled epileptic seizure elicited through a current delivered via electrodes on the skull. ECT has a faster therapeutic onset than antidepressant drugs. Even so, several treatment sessions are required before remission is achieved (Ferrier and Waite, 2020). In terms of efficacy and onset of action, ECT is superior to conventional antidepressants; however, ECT’s main drawback include a number of side effects, such as cognitive impairment, delirium, musculoskeletal pain/injury, and anesthesia-related complications (Andrade, 2022). Some patients remain hesitant about choosing ECT over concerns about cognitive side effects, which are common and can be disturbing but are rarely persistent (Semkovska and McLoughlin, 2010). Furthermore, it carries a well-known stigma among patients and their families (Hay et al., 1989). A scientometric analysis by Tran et al. (2019) concluded that the ECT’s popularity index has been relatively stable over the past few decades, with a slight increase in recent years. The paper analyzed scientometric data from 1988 to 2017 and found that while there has been a decline in the use of ECT in clinical practice, there has also been an increase in the number of scientific publications and citations related to ECT. This suggests that while ECT may not be as commonly used in clinical practice, it continues to be an important and relevant treatment option for some patients with severe mental illness.

There is a rapidly changing picture emerging from previous studies that may provide an entirely new set of potential therapeutic targets. Accumulating data in recent years have demonstrated that the excitatory glutamatergic/N-methyl-d-aspartate receptor signaling may play a pivotal role in the pathophysiology of MDD (Ghasemi et al., 2014). Ketamine is a noncompetitive high-affinity N-methyl-d-aspartate receptor antagonist used for anesthesia, and clinical studies have suggested its rapid-onset antidepressant effects (Fond et al., 2014; Sadock et al., 2015; Zarate et al., 2006).

Berman et al. (2000) performed the first clinical trial that revealed a significant rapid antidepressant effect with an infusion of 0.5 mg/kg of ketamine for 40 min. Additionally, ketamine is comparatively safe in sub-anesthetic doses (0.1–1 mg/kg) except for some transient side effects, including dry mouth, dizziness, blurred vision, headache, nausea or vomiting, restlessness, tachycardia, increased blood pressure, and dissociation/perceptual disturbances (Fond et al., 2014). Ketamine could be an interesting alternative in TRD due to its potential fast onset, impact on suicidal thoughts, multiple routes of administration, and lower cost. However, the long-term side effects of ketamine treatment in depressed patients are only partially known, and concerns for the safety of ketamine as an antidepressant treatment have been raised (Ekstrand et al., 2022). Ketamine may be also associated with misuse for recreational reasons, creating concerns for abuse liability when prescribed for depression. Also, since the 1980–90s, ketamine has gained popularity worldwide as a “club-drug” in association with the youth dance culture (Le et al., 2022). Assessment of history of substance abuse should also be compulsory before any off-label prescription of ketamine is made (Ceban et al., 2021; Zhang et al., 2016). Research indicates that high-frequency ketamine is associated with urothelial barrier disruption, inflammation in the bladder, ketamine direct toxicity, nerve hyperplasia and hypersensitivity, cell apoptosis, microvascular damage, and overexpression of carcinogenic genes (Ng et al., 2021a,b).

Recently, new studies were performed to compare the efficacy between ECT and ketamine through rating scale changes, as following: Hamilton Depression Rating Scale-25 (HDRS25), Beck Depression Inventory (BDI), Beck Scale for Suicidal Ideation (BSSI), and Montgomery–Åsberg Depression Rating Scale (MADRS) and potential side effects. We performed a systematic review and meta-analysis of both randomized controlled trials and observational studies evaluating the efficacy and side effects that they presented, once no extant or ongoing studies have yet performed a statistical analysis comparing the latest published studies about the subject. As they are relatively new studies and few published in the literature, we will take into account the biases presented by them.

2. Methods

2.1. Eligibility criteria

Considering that there are few randomized studies in the literature about the subject, we decided to include both randomized clinical trials and observational studies. Inclusion criteria were the comparison between sub-anesthetic doses (0.1–1 mg/kg) of ketamine (IM, IV) and ECT, electrode placement provided (bitemporal, unilateral, or mixed), subjects ≥18 years old of either sex with a diagnosis of MDD (unipolar or bipolar depression according to the DSM-IV, DSM-V and ICD-10) with or without psychotic symptoms with a clinical indication for ECT and able to provide voluntary consent. Exclusion criteria were uncontrolled severe medical illness, history of neurological disease, substance use dependency (except caffeine and nicotine), cardiovascular diseases in the past six months, untreated hypo or hyperthyroidism, pregnant or breastfeeding, lifetime antidepressant treatment with ketamine, history of recreational use of ketamine or ketamine use as an anesthetic for the ECT group sample.

2.2. Search methods

We searched the bibliographic databases MEDLINE, Web of Science, Embase, PsycINFO, and Google Scholar without restrictions on publication date. We also searched trial registers, which were the Cochrane Register of Controlled Trials, ClinicalTrials.gov, and the World Health Organization’s International Clinical Trials Registry Platform (WHO-ICTRP), to identify any unpublished or ongoing trials. Two authors (AS & GT) independently screened the titles and abstracts of potentially eligible studies and extracted data. Later, another pair of authors (DM & SC) performed the full-text analysis. Interrater reliability was substantial (κ = 0.76) for title/abstract screening and almost perfect (κ = 0.85) for full-text screening. Disagreements were resolved through discussion. The complete search strategies (stratified for each search platform) are available in the Supplementary Material.

2.3. Data acquisition

A data extraction spreadsheet was constructed first to collect general information about the included studies. Data were extracted on: (a) study/context characteristics: authors, journal name, year of publication, inclusion and exclusion criteria, the geographic region where the study was conducted, study design, duration and follow-up times, and methods of assessment of different outcomes; (b) participant aspects: age and sex distributions; (c) interventions’ particularities: dosage and route of administration of ketamine; (d) outcomes: severity of depression through scales, depression response and side effects. Depression response was defined as a reduction in the scores of the scales by 50 % at the end of the treatment or following the last therapy session.

We extracted all adverse events reported. Only endpoints with three or more studies were included in the quantitative analysis. We also extracted arm-level data, including information about sample sizes, observed events, means, standard deviations, results of hypothesis tests, and effect size measures according to each endpoint (when reported) for both compared interventions at the end of the treatments. Hedges’ g and log-corrected relative risks were computed through R version 4.0.5 (Windows), using different packages.

2.4. Risk of bias and quality assessment

The risk of bias in included studies was assessed by two independent raters using the “Cochrane Risk of Bias 2” (RoB 2.0) tool for RCTs and the “Risk of Bias in Non-Randomized Studies” (ROBINS-I). Once studies were determined to fit the inclusion criteria, additional data were extracted for each study to specifically assess for adequate random sequence generation, allocation concealment, subject blinding, outcome blinding, and procedures for dealing with incomplete data and selective reporting.

2.5. Data synthesis

Random-effects models were chosen because the expected heterogeneity among included studies would likely lead to varied true effect size between the units of analysis. By definition, random-effects models attempt to generalize findings beyond the included studies by assuming that the selected studies are random samples from a larger population (Cheung et al., 2012). Statistical heterogeneity was evaluated by I² (with test-based 95 % confidence intervals). In addition, Knapp and Hartung (2003) adjustments were applied. For the continuous endpoint (depression severity through scale) data pooling: (a) we used the restricted maximum likelihood method (Viechtbauer, 2005) to estimate the (heterogeneity) of variance of the distribution of the true effect sizes (τ²); and (b) Hedges g’ corrected standardized mean difference was calculated, since the number of studies included is small. For binary continuous endpoint (response to therapy and adverse effects) data pooling: (a) the Mantel-Haenszel method was selected; (b) we used the Paule-Mandel estimator of τ², and (c) the effect size measure calculated was risk ratio. The pooled effects results with 95 % confidence intervals were presented in different forest plots and tables, sorted by year of publication.

An influential analysis was performed using different approaches, only for pools with more than three studies and moderate or substantial heterogeneity (I² > 50 %). The first involved the detection and removal of outliers. After that, we detected the influential studies using the leave-one-out principle, described using Baujat plots and influence plots.

We chose to perform a subgroup analysis comparing the studies by the degree of risk of bias as an exploratory approach. Although in the registry we pointed out the possibility of dividing by sex, country, type of study, and scale used, we did not obtain enough studies in at least one of the categories of each variable.

During prospective registration, we planned to analyze the risk of publication bias through funnel plots and their respective statistics. However, because of the number of studies included, we decided not to perform this analysis due to its low statistical power (Sterne et al., 2011). Despite this, previous measures were taken, such as a gray literature search, to avoid further problems with selective reporting.

This study was conducted following the Cochrane Handbook and was registered with PROSPERO (registration number CRD42022349220). All reports were consistent with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). All scripts used to perform the models are available in the Supplementary Material.

3. Results

3.1. Literature search

A total of 2875 articles were retrieved. After the full-text screening of 17 studies, eight met the inclusion criteria (Fig. 1). The dosing of ketamine was similar among the studies, although the intramuscular route was used in one study. The studies are further described in Table 1.

Fig. 1.

PRISMA flow diagram of study screening and selection.

Table 1.

Characteristics of studies included in the meta-analysis.

Author, year	Country	Study design	Interventions details		Treatment duration	Follow-up	Number of patients (KET:ECT)	Age (years)	Depression scale(s)	Baseline score of HDRS or MADRS
			KET dose (route, duration)	ECT laterality (anesthetic)						KET; Mean (SD)	ECT; Mean (SD)
Ekstrand et al., 2022	Sweden	RCT	0.5 mg/kg (IV, 40 min)	96 % unilateral (NI)	Trice weekly until remission, maximum 12 sessions	1 week, 3,6−, and 12− month post treatment	95:91	18–85	MADRS	33.1 (6.3)	34.5 (5.7)
Kheirabadi et al., 2020	Iran	RCT	0.5 mg/kg (IM, 40 min)	Bilateral (succinylcholine)	6–9 sessions across 3 weeks	1− and 4− week post treatment	15:12	18–70	HDRS17	21 (2.9)	21.83 (4.63)
Sharma et al., 2020	India	RCT	0.5 mg/kg (IV, 45 min)	Both (thiopentone & succinylcholine)	Alternate days, thrice a week, for 2 weeks. Maximum 12 sessions.	Until the end of treatment	12:13	18–65	HDRS17, BDI	23.33 (4.05)	25.15 (6.58)
Basso et al., 2020	Germany	Cohort	0.5 mg/kg (IV, 40 min)	Unilateral (succinylcholine)	Thrice weekly for 4 weeks	Until the end of treatment	25:24	20–71	MADRS	26.40 (4.94)	31.17 (7.28)
Loureiro et al., 2020	USA	Cohort	0.5 mg/kg (IV, 40 min)	Unilateral, 48 % bilateral (NI)	Twice or thrice weekly for a total of four infusions	Until the end of treatment	27:17	25–48	HDRS17	20.15 (4.7)	21.41 (8.33)
Kheirabadi et al., 2019	Iran	RCT	0.5 mg/kg (IV, 40 min)	Bilateral (thiopental, atropine & succinylcholine)	Twice weekly up to 6 weeks	12 months post treatment	10:12	19–59	HDRS	24.6 (2.4)	26.1 (3.8)
Allen et al., 2015	Ireland	Cohort	0.5 mg/kg (IV, 40 min)	Bitemporal (methohexitone and suxamethonium)	Twice weekly for ECT; up to three infusions once a week for ketamine	Until the end of treatment	17:18	32–53	HDRS17	21.6 (3.09)	20.2 (6.36)
Ghasemi et al., 2014	Iran	RCT	0.5 mg/kg (IV, 45 min)	Bilateral (thiopental, atropine & succinylcholine)	3 sessions every 48 h	1 week	9:9	18–75	HDRS25, BDI	30.22 (5.78)	35.88 (6.47)

NI = not informed; KET = ketamine; ECT = electroconvulsive therapy.

3.2. Quality assessment

Two of the randomized studies had a high risk of bias, while all the observational studies were rated as having serious risk of bias. Representative traffic light plots of the analysis were generated using the weight of the studies on the primary outcome, shown in Fig. 2.

Fig. 2.

Risk of bias assessment of randomized studies (A) and observational study (B). Only the first three letters of each main author are displayed.

3.3. Primary outcome

We were able to use patient-level data or reported values to calculate the standard error in the change of depression scales. We contacted some authors to obtain the data required in cases where the original article did not provide it. Two trials demonstrated statistically significant differences in the reduction of depressive symptoms scales between ketamine and electroconvulsive therapy.

Effect estimates and standard errors were used to perform a random effects meta-analysis of the comparison between intervention groups at the end of the treatment (Fig. 3). The pooled estimate of the Hedges’ g in depression scores between the ketamine and electroconvulsive therapy arms was −0.12 (95 % CI: −0.77 to 0.54), where negative values suggest ketamine was superior to electroconvulsive therapy in reducing depressive symptoms. The between-study heterogeneity variance was estimated at τ²-estimate = 0.36 (95%CI: 0.08 to 2.41), with an I² value of 77 % (95%CI: 53 to 89.2 %).

Fig. 3.

Forest plot of mean change in depression symptoms severity scales for ketamine compared with ECT. IV: intravenous; IM: intramuscular.

Since a high heterogeneity was found, we chose to perform outlier and influence analyses. No outliers were detected through the confidence interval overlap analysis. Further, analysis of the Baujat and covariance ratio influence plots (Fig. 4) indicated the study by Loureiro et al. (2020) as the one contributing the most to heterogeneity. A pooling without this influential study was performed. The results for comparison are shown in Fig. 5. The exploratory results of the subgroup analysis according to the risk of bias are shown in Fig. 6.

Fig. 4.

Influence plots regarding studies of the primary outcome.

Fig. 5.

Forest plot of mean change in depression symptoms severity scales for ketamine compared with ECT without an influential case. IV: intravenous; IM: intramuscular.

Fig. 6.

Forest plot of subgroup analysis through risk of bias comparing ketamine versus ECT regarding the primary outcome. IV: intravenous; IM: intramuscular.

3.4. Secondary outcomes

3.4.1. Response to therapy

A random effects meta-analysis of the comparison of response between the different intervention groups was performed using effect estimates and standard errors (Fig. 7). Only three studies reported this outcome. The pooled risk ratio estimate in response to therapy between treatment groups was 0.89 (95%CI: 0.55 to 1.45), where values between zero and one suggest electroconvulsive therapy was superior to ketamine in promoting response. The between-study heterogeneity variance was estimated at τ²-estimate = 0.04 (95%CI: 0 to 1.38), with an I² value of 45.8 % (95%CI: 0 to 82 %).

Fig. 7.

Forest plot of response to therapy for ketamine compared with electroconvulsive therapy. IV: intravenous.

3.4.2. Adverse effects

Random effects meta-analyses comparing reported side effects between ketamine versus ECT intervention groups were performed using effect estimates and standard errors. Although numerous adverse events were reported in most of the included studies, only three of them collectively reported the following adverse effects during the trial: dissociative symptoms, nausea, muscle pain, and headache. The pooled estimate of the risk ratio for dissociative symptoms was 5.42 (95%CI: 0.72 to 40.87), and the between-study heterogeneity variance was estimated at τ²-estimate = 0.20 (95%CI: 0 to 36.14), with an I² value of 22.9 % (95%CI: 0 to 92 %). The pooled estimate of risk ratio for nausea was 0.73 (95%CI: 0.16 to 3.44), with a τ²-estimate = 0.14 (95%CI: 0 to 19.22) and an I² value of 44.3 % (95%CI: 0 to 83.4 %). The pooled estimate of risk ratio for muscle pain was 0.25 (95%CI: 0.10 to 0.65), with a τ²-estimate = 0 (95%CI: 0 to 33.95) and an I² value of 0 % (95%CI: 0 to 89.6 %), as shown in Fig. 4B. Finally, the pooled estimate of risk ratio for headache was 0.39 (95%CI: 0.11 to 1.35), with τ²-estimate = 0.12 (95%CI: 0 to 8.76), with an I² value of 70.1 % (95%CI: 0 to 91.2 %). Despite the high heterogeneity in the analysis of some secondary outcomes, we chose not to perform an influence assessment due to the low number of studies. Forest plots related to those analyses are depicted below (Figs. 8–11).

Fig. 8.

Forest plot comparing ketamine versus ECT regarding dissociative symptoms. IV: intravenous; IM: intramuscular.

Fig. 11.

Forest plot comparing ketamine versus ECT regarding muscle pain. IV: intravenous; IM: intramuscular.

4. Discussion

To the best of our knowledge, this is the first meta-analysis to investigate the efficacy of ketamine compared with ECT in the treatment of patients with TRD. The main findings of our meta-analyses were as follows: (a) there was no statistically significant difference between the groups regarding the improvement of depressive symptoms (primary outcome); (b) despite the depression response outcome having favored ECT, the comparison between the two groups was not statistically significant, therefore it is not possible to conclude that ECT is superior to ketamine regarding response rates; (c) muscle pain was the only side effect with statistical significance, favoring ketamine over ECT. Then, we can conclude that there is a lower risk of having muscle pain as a side effect when using ketamine compared with ECT.

We performed an analysis of influence between the studies concerning the primary outcome of our investigation which showed that Loureiro et al. (2020) contributed the most to the heterogeneity in the data. Participants from that study were assigned to either ketamine or ECT treatment. However, patients in the ECT group received a greater number of sessions (14 on average), while participants in the ketamine arm received only 4 sessions in total. The authors chose to maintain patients under the intervention in the ECT group until participants achieved maximal response or remission for at least a week. Moreover, ECT-treated patients were not allowed to maintain their antidepressant medications prior to treatment, whereas the ketamine participants were allowed to continue on their usual antidepressant drug regimen. Also, all patients had moderate to severe disease with HDRS scores ≥18. These may be important reasons for its contribution to the heterogeneity of the analysis. Nevertheless, despite removing this study in the sensitivity analysis we still found high heterogeneity. Upon evaluation of the research design of the units of analysis, we observed that two studies had methodologies that were interestingly different from the others. Participants from Kheirabadi et al. (2020) were diagnosed with MDD by the DSM-V criteria, referred by their psychiatrist for ECT, and had baseline HDRS17 scores of 21 ± 2.9 (ketamine) and 21.83 ± 4.63 (ECT). In addition, this study compared IM and oral ketamine with ECT. For the analysis, we decided to consider the results of IM ketamine intervention due to equivalent bioavailability to IV administration. The other study that showed high methodological heterogeneity (Ghasemi et al., 2014) had patients with a major depressive episode, scheduled to receive ECT, with a diagnosis according to the DSM-IV and baseline HDRS25 scores of 30.22 ± 5.78 (ketamine) and 35.88 ± 6.47 (ECT). The mean baseline scores from participants of both studies suggest that most patients had severe disease at the beginning of treatment (Blacker, 2005). Such sample characteristics may also have contributed to the discrepancy and increased levels of heterogeneity in the analysis. The study by Ekstrand et al. (2022), which had the broadest number of participants and weighted the most in the results, selected participants with unipolar depression who were hospitalized and diagnosed according to the DSM-IV with a score of ≥20 on the MADRS. Mean baseline participants’ scores were 33.1 ± 6.3 (ketamine) and 34.5 ± 5.7 (ECT), which suggests that most patients in this study had a moderate depression severity, taking into account the classification of Mullerthomsen et al. (2005).

Until now, ECT is the most effective treatment for severe depression and TRD. The mechanism of action of ECT has not yet been fully understood (Leiknes et al., 2012), which can contribute to its stigma, despite its proven potential efficacy and safety. In addition, no treatment has achieved equivalent or greater efficacy than ECT for patients with “severe mood” and “psychotic disorders” (Kellner et al., 2020; Tørring et al., 2017). Despite the successful remission rates of ECT, one-third of patients still do not respond. The possibility of finding another treatment that could be as effective and safe as ECT would represent an additional therapeutic tool against TRD and could bring relief to those patients to whom most treatments have failed. Due to its rapid onset and antidepressant effect, ketamine could be an alternative treatment. Several studies since the 90s have demonstrated the antidepressant effect of ketamine, especially concerning patients with MDD and TRD (aan het Rot et al., 2010; Berman et al., 2000; DiazGranados et al., 2010; Murrough et al., 2011; Thakurta et al., 2012; Trullas and Skolnick, 1990).

A meta-analysis by McIntyre et al. (2020) reviewed randomized, double blind, placebo-controlled trials on the effects of the drug by different routes of administration (IV, IN and oral) in mood disorders, concluding that ketamine had significant antidepressant effects in all of the formulations, especially when administered via the IV or the intranasal (IN) routes. IV ketamine reduced depression scores the most within the first 2–6 days of treatment, while IN ketamine appeared to be more effective in the first 24 h. Oral intake, yet efficient, appeared to reproduce less robust results and had more issues regarding heterogeneity and number of available studies. Nonetheless, the efficacy of the drug in the short-term when administered IV or IN was demonstrated. However, a meta-analysis by Alnefeesi et al. (2022), which focused on naturalistic studies in order to establish real-world clinical effectiveness of the drug for patients with TRD, discussed that ketamine could actually be a viable mid-to-long term therapy in some patients as meta-regressions confirmed that dose frequency, dose number, and the duration until follow-up had no effect on any treatment outcomes. The authors also found substantial antidepressant effects of ketamine, although considerable variation of effectiveness was noted across the different clinical populations.

In order to better establish the effectiveness of ketamine for patients with TRD, Rodrigues et al. (2022) compared the use of intravenous ketamine in patients that had a history of neurostimulation with those that had never been exposed to neither ECT nor repetitive transcranial magnetic stimulation, considering that the population exposed to neurostimulation therapies may be a particularly resistant set of patients to depression treatments. Both demonstrated moderate to large effects in reduction of depressive symptoms although the results pointed towards no significant differences between the neurostimulation-naïve cohort and the group exposed to such therapies.

Given ketamine’s singular mechanism of action on the glutamate system instead of the monoamine system and its unclear effects in certain domains of depression, besides observing the effects on depressive symptoms improvements some authors also approached other aspects of depression treatment with ketamine. Notably, a study by Di Vincenzo et al. (2022) evaluated the frequency in which patients under repeated ketamine infusions experienced depressive symptoms worsening in a naturalistic, real-world sample. The findings suggested that those under treatment with the drug presented with similar rates of symptomatic worsening of other pharmacological and non-pharmacological treatments.

The effects of ketamine on specific domains of depression have also been increasingly commented on in newer studies. Jawad et al. (2023) conducted a review which concluded that both ketamine and esketamine treatment improved major psychopathological domains of depression such as anhedonia, cognition, functioning and especially suicidality. Another review, however, by Ng et al. (2021a,b) centered specifically around the effects of ketamine to general functioning (social, workplace, psychosocial) in patients with MDD or TRD had mixed results and the study was met with limitations regarding paucity of literature on the subject, measurements of general functioning as a secondary outcome, and incomplete statistical testing. The reduction in suicide ideation, in particular, has been more thoroughly explored in more recent works. A review by Siegel et al. (2021) commented on RCTs, retrospective open-label trials and case studies about the topic and noticed mixed results as well concerning antisuicidal effects. Interestingly, most of the selected studies reported significant improvement of the depressive symptoms after treatment, even when suicide ideation did not improve as much as with placebo, suggesting that possibly the antisuicidal effects of ketamine are independent from its antidepressant potential.

An RCT evaluating 67 patients with recurrent depression receiving IV ketamine in comparison with placebo exhibited significant improvement in standardized depression scores and response rates, maintaining antidepressant efficacy over 15 days by the ketamine group (Singh et al., 2016). Of note, given its cardiovascular stimulation effect, anesthetic and subanesthetic ketamine administration have been associated with rises in blood pressure and heart rate (Goddard et al., 2021; Vankawala et al., 2021). Nevertheless, only four articles in this meta-analysis mentioned increases in heart rate or blood pressure, which were predominantly transient and few compared to the other adverse effects.

Despite none of the selected articles for this meta-analysis have mentioned subjects with a history of ketamine abuse/dependence (as it was a prohibiting criterion in some), it is a concerning side effect that may be explained by its inherent amphetamine-like properties (Zhang and Ho, 2015). More recent studies have demonstrated how ketamine successfully reverses hypodopaminergic states following acute withdrawal from amphetamine, which does imply that ketamine has inherent stimulant properties and it might cause a rapid improvement in mood, but also cause further addiction issues (Zhang and Ho, 2016).

It is recommended that physicians screen patients carefully for medical and psychiatric conditions that may increase the risk of complications and use a standardized assessment tool to evaluate patients’ readiness for treatment. Clinicians must inform their patients of both the positive and negative aspects of ketamine. Assessment of history of substance abuse should also be compulsory before any off-label prescription of ketamine is made (Ceban et al., 2021; Zhang et al., 2016).

Additional adverse effects related to ketamine therapy include neuropsychiatric: transient dissociation, perceptual disturbances, abnormal sensations, derealization, and also depersonalization; psychotomimetic: induction of psychosis (especially in individuals with related pre-existing conditions (for example, schizophrenia)); psychiatric: anxiety is the most commonly reported adverse effect associated with ketamine; neurologic: limited dizziness/vertigo, drowsiness, sedation, light-headedness, headaches, and blurred vision; cardiovascular: ketamine exert sympathomimetic effects on the cardiovascular system, and patients can experience transient hypertension and tachycardia; gastrointestinal: transient nausea, emesis and loss of appetite (Ceban et al., 2021).

Ketamine-induced uropathy in patients exposed to the drug, especially those with a history of abuse or chronic exposure to ketamine, has also been described in some papers, (Chang et al., 2012; Lai et al., 2012; Chu et al., 2008; Shahani et al., 2007). Available research indicates that high-frequency ketamine is associated with urothelial barrier disruption, inflammation in the bladder, ketamine direct toxicity, nerve hyperplasia and hypersensitivity, cell apoptosis, microvascular damage, and overexpression of carcinogenic genes (Ng et al., 2021a,b). There was also only one mention of urologic complications in one participant from Ekstrand et al. (2022), who reported urinary retention. Interestingly, a systematic review (Castellani et al., 2020) observed several studies that suggested variable urologic damage associated with ketamine use, but mainly abuse, such as hydronephrosis, urinary retention, ureteral stenosis, and bladder fibrosis, which could be explained by direct toxicity to the urothelium by ketamine ions excreted in the urine.

Two studies (Basso et al., 2020; Ghasemi et al., 2014) included in this review (one of which was an observational study) suggested that ketamine reaches antidepressant levels faster than ECT. Two other trials (Ekstrand et al., 2022; Sharma et al., 2020) concluded that ketamine was inferior to ECT, considering remission rates and reduction of depressive symptoms. In contrast, other included studies (Allen et al., 2015; Kheirabadi et al., 2019; Loureiro et al., 2020; Kheirabadi et al., 2020) concludes that ketamine and ECT have similar antidepressant effects, but Kheirabadi et al. (2020) and Kheirabadi et al. (2019) suggests that therapy with ketamine has shown fewer cognitive adverse effects. These studies, however, presented some limitations, as shown in the risk of bias analysis. Five (Ghasemi et al., 2014; Kheirabadi et al., 2020; Kheirabadi et al., 2019; Sharma et al., 2020; Allen et al., 2015) out of eight of the reviewed studies were very small sample sizes. Two of the included studies (Kheirabadi et al., 2020; Kheirabadi et al., 2019) could have therapeutic-related biases due to the suboptimal frequency of the ECT sessions (twice weekly instead of thrice). Besides this, both studies performed a completer analysis instead of an intention to treat. Moreover, Ghasemi et al. (2014) had one MDD patient with comorbid bipolar disorder, which has conflicting diagnostic criteria.

A systematic review performed by Veraart et al. (2021) analyzed some of our included studies (they did not include Ekstrand et al., 2022, and Kheirabadi et al., 2020). They concluded that some of those studies presented with several limitations, including: Allen et al. (2015), that included three patients with history of substance abuse, which could have influenced prognosis and diagnosis, also there is little information about patient selection, and the follow-up period was short (1 week). Additionally, in the ketamine group, there was a 1-week interval between ketamine infusions, which may have negatively influenced antidepressant treatment potential, given that usual dosing is 2–3 infusions per week. The studies of Ekstrand et al. (2022) and Allen et al. (2015) both presented a significant loss of follow-up from participants in the ketamine group (22 % of Ekstrand et al., 2022 and 41 % of Allen et al., 2015), resulting in fewer treatment sessions received on average for this intervention group. Moreover, Allen et al. (2015) and Loureiro et al. (2020) studies were not designed to compare clinical outcomes of the two treatments, but rather to investigate the impact on brain-derived neurotrophic factor and modulation of amygdala reactivity following ECT or ketamine interventions respectively.

4.1. Limitations

Besides the reduced number of available studies, other factors contributed to the limitations of our findings. Some source material had methodological issues, as noted by Andrade (2022), such as poor reporting of data, lack of an intention to treat analysis and contradicting statements. Five out of eight studies were classified as having a serious to high risk of bias, possibly leading to a prejudiced analysis of combined data. Besides, most of them had small sample sizes and presented high in-between heterogeneity, which compromised the quality of the evidence and its applicability to the general population. Even though gray literature was accounted for in the literature review, the effect of publication bias might have limited the information available for this analysis. Contribution to this systematic review came mainly from standard bibliographic databases such as Embase, MEDLINE, and Google Scholar, without any studies found exclusively in other sources.

5. Conclusion

This analysis showed no evidence to support the superiority of ketamine over ECT. Regarding side effects, the evidence suggested a decreased risk of muscle pain in patients treated with ketamine. However, in most studies, this group found drop-outs due to adverse effects more common.

The current available studies are insufficient to generate statistically significant results on primary and most of the secondary outcomes. Further research is needed to determine the superiority of ECT or ketamine in reducing depressive symptoms severity and tolerability profile. Another future meta-analysis could present clearer evidence as a more significant number of studies are being conducted on the subject.

Fig. 9.

Forest plot comparing ketamine versus ECT regarding nausea. IV: intravenous; IM: intramuscular.

Fig. 10.

Forest plot comparing ketamine versus ECT regarding headache. IV: intravenous; IM: intramuscular.

Abbreviations:

ECT

Electroconvulsive therapy

MDD

Major depressive disorder

TRD

Treatment-resistant depression

HDRS25

Hamilton Depression Rating Sacle-25

BDI

Beck Depression Inventory

BSSI

Beck Scale for Suicidal Ideation

MADRS

Montgomery-Åsberg Depression Rating Scale

ICD

International Statistical Classification of Diseases

DSM

Diagnostic and Statistical Manual of Mental Disorders