Ketamine Versus Electroconvulsive Therapy for the Treatment of Depression: A Guide for Clinicians

Abstract

The effective treatment of major depressive disorder remains one of the biggest public health challenges globally. For moderate to severe cases, pharmacotherapy often falls short, leading to treatment-resistant depression. Electroconvulsive therapy (ECT) has generally been considered the gold standard for severe cases of treatment-resistant depression. However, emerging evidence suggests that ketamine may serve as a promising alternative. Two relatively large noninferiority trials and three meta-analyses support the efficacy of both treatments but report contradictory findings regarding superiority. The authors discuss possible reasons underlying these discrepant findings, including variations in patient selection criteria, study outcome measures, treatment delivery, and site experience. Additionally, the authors examine the unique risk and benefit profiles of each treatment, highlighting patient-specific considerations. By evaluating the most recent evidence for the efficacy of ketamine versus ECT alongside key patient-specific factors, the authors aimed to guide clinicians in recommending the optimal treatment choice for each patient.

The pervasive reach and impact of depression is well known. It affects 280 million people worldwide, it has a severe impact on quality of life and functioning, and it is associated with a significant economic burden (1). For moderate to severe major depressive disorder, pharmacotherapy is the standard first line of treatment, although antidepressants are limited in efficacy for many, and the onset of therapeutic effects can take several weeks. Consequently, approximately 30% of patients with major depressive disorder meet the criteria for treatment-resistant depression (TRD), defined as the failure to respond to at least two antidepressant trials (2). Moreover, evidence suggests that as the number of failed antidepressant trials increase, response and remission rates decrease (3).

People with TRD—especially those with high-severity TRD—can benefit significantly from augmentation with interventional treatments such as electroconvulsive therapy (ECT). ECT has been used in psychiatry for over 80 years, first receiving approval as a Class-III (high-risk) medical device in 1976 (4). In 2018, the U.S. Food and Drug Administration (FDA) reclassified the use of ECT for TRD from Class III to Class II (moderate risk, requiring special controls) (5). It has rapid onset effects that make it a favorable choice for patients who are severely depressed. Despite its effectiveness and the use of contemporary techniques that have improved the side-effect profile, the use of ECT has been declining in the United States in recent decades, which is attributed in part to negative stigma, the possibility of associated cognitive side effects, and cost (6).

More recently, ketamine—an N-methyl-d-aspartate receptor antagonist that was originally approved as an anesthetic—has emerged as a rapid-acting antidepressant. In 2000, Berman et al. (7) conducted a placebo-controlled, double-blind, crossover trial in which participants received a single subanesthetic dose of ketamine or saline placebo. Depressive symptoms were significantly decreased within 72 hours of administration of ketamine but not saline placebo (delivered over 40 min) (7). After subsequent industry-sponsored trials, esketamine, the S-enantiomer of ketamine in the form of a nasal spray (Spravato) received FDA approval for the treatment of TRD in 2019, with a second indication granted in 2020 for the treatment of major depressive disorder with suicidal ideation or behavior.

ECT has long been considered the gold standard for treating adults with TRD. However, the mounting evidence of the efficacy of ketamine and esketamine in patients with severe TRD has led many to question when ECT or ketamine should be considered in the treatment algorithm, and for which patients. This review evaluates the clinical efficacy and distinct safety profiles of ECT and of ketamine and esketamine for the treatment of unipolar major depressive disorder and outlines important patient-specific considerations to provide evidence-based guidance for clinicians on which intervention to choose for their patients.

Ketamine Versus ECT: Comparison Trials and Meta-Analyses

Noninferiority Trials

Two large noninferiority trials have examined the comparative efficacy of ketamine and ECT in adults with major depressive disorder to date. The first trial, called “KetECT,” enrolled 186 Swedish patients who were hospitalized with major depressive disorder with or without psychotic features (ages 18–85 years) (8). Participants were randomized to receive either ECT or ketamine infusions (0.5 mg/kg/40 min) three times per week until remission or a maximum of 12 treatment sessions. The primary outcome was remission according to the Montgomery-Åsberg Depression Rating Scale (MADRS), and remitters were followed for 12 months. ECT demonstrated significantly higher remission rates, with 63% of the patients who received ECT remitting versus 46% of the patients who received ketamine. In the 12-month follow-up period, 64% of the remitters in the ECT group and 70% of remitters in the ketamine group relapsed, with no significant difference in time to relapse.

In the KetECT trial, the researchers observed distinct patterns of adverse events (AEs) between treatments. The most common AEs associated with ECT were headache, muscle pain, and amnesia, whereas ketamine recipients more often reported dissociative side effects, anxiety, blurred vision, euphoria, vertigo, and diplopia. ECT recipients experienced side effects that lasted 24 hours or longer twice as often as those in the ketamine group, and 21 ECT recipients reported prolonged amnesia and muscle pain, in some cases lasting between 3 and 6 months (N=3) or through the end of follow-up (N=3). Twenty-six percent of patients in the ECT group reported a severe AE, compared with 15% in the ketamine group.

The second and largest noninferiority trial to date, called “ELEKT-D,” enrolled 403 participants (ages 21–75 years) with nonpsychotic TRD, primarily from outpatient settings at five academic sites within the United States (9). Participants were randomized to receive either ECT three times per week or ketamine infusions (0.5 mg/kg/40 min) twice per week for 3 weeks. The primary outcome was treatment response (a decrease of at least 50% from the baseline score) as measured by the 16-item Quick Inventory of Depressive Symptomatology–Self-Report (QIDS-SR-16). A response as determined with the QIDS-SR-16 occurred in 55% of patients in the ketamine group and 41% in the ECT group, showing noninferiority of ketamine to ECT. Remission according to the clinician-reported MADRS, a secondary outcome in the study, was 37.9% in the ketamine group and 21.8% in the ECT group. Patients who met the response criteria were followed over 6 months. Relapse, defined as a QIDS-SR-16 score higher than 11, was measured at 1, 3, and 6 months. Relapse occurred in 19% of patients in the ketamine group and 35% of patients in the ECT group at 1 month; in 25% and 51%, respectively, at 3 months; and in 34% and 56%, respectively, at 6 months.

In the ELEKT-D trial, a higher percentage of participants who received ECT reported musculoskeletal AEs, and scores of dissociative symptoms were higher among participants who received ketamine. Thirty-two percent of ECT recipients had at least one moderate or severe AE, compared with 25% of ketamine recipients.

In summary, the two trials drew contradictory conclusions, with KetECT deeming ECT superior and ELEKT-D deeming ketamine noninferior and, in fact, showing ketamine to outperform ECT numerically on almost all measures. Several factors may be responsible for the differences in the study results, including patient selection and characteristics, choice of outcome measures, differences in treatment delivery, and site experience and/or clinician comfort with the two treatments. Each of these potential explanatory factors are discussed in the following text.

Patient selection and characteristics.

KetECT enrolled 100% of the participants from inpatient settings. The mean duration of the current depressive episode was 3.2 months, 17% had depression with psychotic features, and the mean age was 52 years. Contrarily, ELEKT-D enrolled 89% of the participants from outpatient settings; patients had a longer mean duration of current depressive episode of 24 months, 0% exhibited psychotic features, as it was an exclusion criterion, and the mean age was 46 years. It is important to note that inpatient hospitalization, shorter duration of major depressive episode, depression with psychotic features, and older age are all positive prognosticators of ECT outcomes (10), possibly contributing to the superiority of ECT in KetECT but not in ELEKT-D.

Choice of outcome measures.

The primary outcome measures differed: KetECT used remission by the clinician-administered MADRS (a score of ≤10) and ELEKT-D used response by the patient-reported QIDS-SR-16 (a decrease of ≥50% from baseline). Given that response is less difficult to attain than remission, this may have contributed to the noninferiority finding in ELEKT-D. However, even on measures of remission, ketamine numerically outperformed ECT, although ELEKT-D’s rates of remission as determined by the QIDS-SR-16 and the MADRS, secondary outcomes, were markedly lower than KetECT’s rates of remission as determined by the MADRS in both treatment groups.

Differences in treatment delivery.

The studies differed in their administration of the treatments. In the KetECT trial, ketamine was administered at 0.5 mg/kg/40 min for a maximum of 12 sessions, whereas in the ELEKT-D trial, ketamine was administered at 0.5 mg/kg/40 min for a maximum of six sessions, with allowance for dose adjustment if clinically necessary, although nearly all patients received 0.5 mg/kg/40 min for all six treatment sessions. The additional sessions in KetECT may have contributed to the higher remission rates seen in the ketamine group. Of note, remission and response rates to ketamine in ELEKT-D are comparable with those seen in real-world outcome data of ketamine administered intravenously (IV ketamine; 54% and 29%, respectively) (11). For ECT, KetECT used a right unilateral, brief-pulse protocol administered three times per week for 4 weeks, and 9% of the patients received bilateral ECT. ELEKT-D used a right unilateral, ultrabrief-pulse protocol three times per week for 3 weeks, with 39% of the patients switching to bilateral lead placement if they had an inadequate response. However, given the limit of nine treatments, many patients received only a few bilateral ECT treatments under the study protocol. The percentage of patients in ELEKT-D who responded or remitted to ECT was lower than in other reports (12). It is possible that, if ECT was started with brief-pulse treatment or allowed for more bilateral lead placement treatments, and patients received a greater number of overall sessions, then higher response and remission rates would have been observed.

Site experience and comfort.

KetECT observed a higher dropout rate in the ketamine group (22%) than in the ECT group (4%) after beginning treatment. The authors primarily attributed this to patient discomfort from ketamine’s dissociative side effects. They also noted that participating sites had extensive experience with ECT but no experience conducting ketamine treatment before study commencement and that the sites were less prepared for the adverse psychological effects of ketamine. The ELEKT-D study clinicians had extensive expertise in both ECT and ketamine delivery. In ELEKT-D, there were near-identical rates of patients completing the defined number of treatments once treatment was started (93% and 92%, respectively). It is possible that, if the clinicians who administered ketamine in KetECT had the level of ketamine-specific experience of ELEKT-D’s clinicians, they may have observed lower dropout rates in the ketamine group and higher rates of remission as a result of more treatments.

Meta-Analyses

Three meta-analyses have compared ECT and ketamine. Rhee et al.’s meta-analysis (13) included six short-term studies (pooled N=340) of various sizes and quality levels, including KetECT but not ELEKT-D. The primary outcome was improvement in depressive symptoms, and secondary outcomes included improvements in trajectory of suicidal ideation, cognition and memory, and safety. ECT was more efficacious than ketamine in reducing depression severity. However, the authors note that the individual needs of patients should be considered, as three studies (14–16) reported that ketamine had more rapid antidepressant effects. Two studies (14, 16) examined cognitive outcomes, but because of limited power, Rhee et al.’s meta-analysis could not draw definitive conclusions about cognition and memory performance. Four studies reported AEs, and ketamine was shown to have a lower risk for headache and muscle pain, whereas ECT had a lower risk for blurred vision, diplopia, vertigo, or dissociation, with no group difference in risk of nausea. One study reported serious AEs, but the number was not statistically significant between groups (8).

Menon et al. (17) examined five randomized controlled trials (RCTs; pooled N=278), including KetECT, but excluded three in their sensitivity analysis because of concerns about study methods and reporting quality. Their primary outcome was improvement in depression outcomes. In the analysis of five RCTs, there was no statistically significant difference in depression scores at posttreatment between ketamine and ECT. However, after the sensitivity analysis, ECT was associated with significantly superior rates of response and remission, with the number of sessions to achieve response and remission also favoring ECT. That said, the authors suggested that the advantage of ECT over ketamine may be smaller than Rhee et al. (13) showed and that ketamine may be worth considering before a trial of ECT for patients who want to be protected from cognitive risks, although posttreatment cognition scores showed little difference between groups.

The most recent meta-analysis by Petrucci et al. (18) of six RCTs (pooled N=655) has built on the prior meta-analyses by including ELEKT-D data. The primary outcome was response rate. Secondary outcomes were remission rates; relapse rates at 1, 3, and 6 months; posttreatment cognition; and safety. The presence of psychotic features, admission status, and age were examined in subanalyses. Rates of response did not differ significantly between the ketamine and ECT groups. When limited to inpatient populations, ECT was superior to ketamine in both response and remission rates. Only one study (9) compared ketamine and ECT in the outpatient setting, and it found that ketamine was noninferior, implicating ketamine as a viable alternative to ECT for outpatients. Relapse rates were similar between groups at all follow-up timepoints. Ketamine was associated with better posttreatment cognition scores than ECT, and the authors noted that this possible protective effect should be considered in the treatment decision. Ketamine had a lower risk of temporary muscle pain, and more common incidence of transient dissociation-depersonalization symptoms. There was no difference in the incidence of headaches. Only one study (8) stratified the results of patients with psychotic features; therefore, the comparative efficacy in this subgroup could not be analyzed. There was no significant interaction between mean age and response. Two of the studies (8, 9) suggested an advantage of ketamine in younger patients, but these subanalyses were likely underpowered.

There are several limitations in all three meta-analyses. Most of the power for the first two meta-analyses came from KetECT, because the other studies had relatively small sample sizes. Inclusion and exclusion criteria varied among the studies, such as the inclusion of bipolar depression, the presence of psychotic features, and hospitalization status. Studies implemented different treatment protocols in the administration of ketamine and especially of ECT. Many of the studies had a high risk of bias and lacked long-term follow-up. In a real-world clinical setting, both ECT and ketamine typically require continuation and maintenance treatments after an acute treatment series.

Regarding the implications on longer-term treatment, we now have data that demonstrate the long-term efficacy of esketamine treatment administered intranasally (IN esketamine) for relapse prevention, lasting up to 4.5 years, with no new safety signals identified (19). However, the aforementioned studies investigated the comparative efficacy of IV ketamine and ECT, not IN esketamine and ECT. Given the lack of data that rigorously compare IV ketamine with IN esketamine for the treatment of major depressive disorder, and the possible differential effects that may exist between the two treatments, we should remain very cautious about inferring any comparisons for either short-term or longer-term treatment between IN esketamine and IV ketamine or ECT treatments.

Maintenance ketamine has shown effectiveness in sustaining antidepressant response, but rigorous studies are lacking (20). Studies of continuation and maintenance ECT show efficacy in preventing relapse compared with pharmacotherapy alone, without a decrease in global cognitive function (21). Long-term follow-up studies are needed to shed light on the comparative durability of response and remission, as well as the tolerability of repeated treatments. Additional strategies to prolong antidepressant effects (e.g., with psychotherapy or concomitant medication) also need to be further investigated.

Another important consideration involves the issues of expectations and contextual effects. Open-label studies, in which participants know whether they were receiving ketamine or ECT, risk significant bias because of the lack of blinding. In RCTs, double blinding is the gold standard; however, in these studies, blinding patients to treatment allocation was challenging because of marked differences in the delivery of each intervention. This limits the ability to control for bias introduced by patient and clinician expectations, which can contribute to placebo and nocebo effects (22). The placebo effect can enhance perceived treatment efficacy, whereas the nocebo effect can worsen outcomes or induce AEs. For example, patients who are aware that they are receiving ketamine, which has garnered significant media attention as a rapid-acting antidepressant (23), might respond better because of their positive expectations. Conversely, those who received ECT, which carries stigma and concerns about cognition side effects in the media (24), might be predisposed to nocebo effects, potentially reducing perceived efficacy or increasing reported AEs. It is also possible that patients and clinicians may have been more comfortable and had more positive expectations with ECT treatment, considering its long track record and reputation as the gold standard in treating previously nonresponsive depressive episodes. One way to control for these effects in future studies that carry the risk of unblinding is to investigate patients’ expectations and preferences before randomization. Patient expectations should also be monitored throughout the study.

At present, it remains unclear whether ketamine is noninferior to ECT for treatment of depression, as illustrated in the contradictory findings of the two large noninferiority trials. The existing data point to ECT being superior among inpatients and possibly in patients exhibiting psychotic features. That said, in the head-to-head trials, ketamine demonstrated impressive effectiveness even in these subgroups. Perhaps the critical question is not whether ECT is superior to ketamine or whether ketamine is noninferior to ECT, but what treatment is most appropriate for the specific patient in consideration and what logic should be used to choose between ketamine or ECT for individual patients. This is particularly important given the differing side-effect profiles, safety considerations, accessibility, and patient preferences that exist between the two treatments. These patient-specific considerations will be discussed in the remainder of this review and are summarized in Box 1. We include comments on both ketamine and esketamine treatment, and the reader should make note of which version is indicated in each scenario.

BOX 1. Important patient-specific considerations

Hospitalization Status

• Electroconvulsive therapy (ECT) may be more effective for inpatient populations than ketamine on the basis of the head-to-head trials.

Speed of Onset

• Ketamine has been shown to rapidly induce antidepressant effects. However, in the two main noninferiority studies, there was no convincing difference in onset of antidepressant effects.

• Notably, both ketamine and ECT commonly require continuation and maintenance treatments after an initial treatment phase for relapse prevention.

Cognitive Deficits

• ECT’s short-term cognitive side effects can be distressing to some patients. Individual patient factors that may increase the risk of cognitive effects associated with ECT, along with right unilateral lead placement and brief or ultrabrief pulse widths, should be considered. The relative risk of long-term cognitive effects with ECT remains a matter of debate.

• Recreational ketamine has been linked to cognitive impairment, but there is no clear evidence of measurable sustained deleterious effects when used according to consensus recommendations or FDA REMS protocol. In fact, the treatments have been associated with improvement on several cognitive measures.

Cardiovascular Conditions

• Caution should be taken with both treatments, because transient increases in blood pressure and heart rate can be significant.

• Preexisting cardiac disease, electrode placement, anesthetic agents, number of stimulations, age, and psychotropic drugs are factors that can contribute to cardiac adverse events with ECT.

• Uncontrolled blood pressure should be managed before starting ketamine treatment; and in conditions where an increase in blood pressure risks complications, ketamine should be avoided if possible.

• History of intracerebral hemorrhage or aneurysmal vascular disease are absolute contraindications for esketamine, but not ECT, although care is warranted with ECT.

Psychotic Versus Nonpsychotic Features

• Current data favor ECT for depression with psychotic features. However, more data are needed for ketamine and esketamine.

Substance Use Disorder

• Ketamine has the potential for abuse and misuse and should be administered under clinical supervision, with treatments limited to the number needed to achieve and sustain a response. Although there is a clear theoretical concern, current data are insufficient to determine whether initiating ketamine treatment for depression increases the risk of substance use disorders or of relapse or worsening of existing substance use disorders.

• There is scant evidence to suggest that ECT increases the risk of substance abuse.

Age

• Older age is often deemed a predictor of increased response to ECT.

• Recent studies suggest ketamine and esketamine are well tolerated and can be effective in older adults.

Patient Preference

• Allowing the patient to weigh the comparative risks of each treatment is a productive way to reconcile the distinct side-effect profiles and influences treatment outcome.

Accessibility and Cost

• ECT is commonly limited to the inpatient setting or to those with some form of assistance at home, and its availability is geographically variable.

• The availability of ketamine and esketamine are also geographically variable and requires assistance with transportation.

• Both treatments carry large financial burdens for uninsured patients. ECT is widely covered by most plans, whereas coverage for ketamine and esketamine (especially IV ketamine) is much more limited.

Important Patient-Specific Considerations

Transient cognitive and perceptual changes.

Ketamine, administered at subanesthetic doses similar to 0.5 mg/kg/40 min, has well-known transient effects on perceptual functions, including feelings of dissociation (loosely defined as detachment from reality) (25). Symptoms of dissociation are typically transient, peaking within 1 hour of ketamine administration and resolving completely by 2 hours posttreatment (26, 27). Although typically mild and well tolerated, these side effects can be distressing and should be closely monitored by a health care provider during and after treatment. A monitoring period of 2 hours is currently required by the Risk Evaluation and Mitigation Strategy (REMS) associated with IN esketamine.

Disorientation to place and time immediately after ECT is well documented and is typically transient (28). Posttreatment delirium, a more acute state of confusion, is relatively common in older patients and in patients with a comorbid neurological condition (28). Post-ECT disorientation and delirium can be mitigated by electrode placement, dosage, and individual patient factors (29).

Cognitive impairment.

Many patients develop short-term cognitive deficits after a course of ECT. Although the nature and degree of cognitive deficits vary, research has shown transient deficits in the domains of memory, attention, new learning, executive function, and processing speed immediately after ECT (28, 30, 31). The longer-term effects of ECT on cognition remain a matter of debate, with some studies showing no or minimal long-term cognitive deficits (32) and others showing persistent cognitive changes after ECT (30, 33).

Inconsistencies in the literature likely stem from differences in ECT techniques and cognitive testing across studies. There is a greater association between cognitive side effects and bilateral electrode placement, larger number of treatment sessions, and a higher stimulus charge. Fewer cognitive side effects are associated with right unilateral electrode placement, ultrabrief pulse width, and shorter courses of treatment (28, 33).

Individual patient factors may increase the risk of cognitive deficits after ECT. Patients with low baseline scores on the Modified Mini-Mental State Examination, elderly individuals, those with significant brain injury, those with low intellectual ability, and patients receiving concomitant lithium medication are associated with higher risk of cognitive deficits after ECT (28, 31).

Although ketamine induces deficits in multiple cognitive domains within the 2-hour period after administration (34), and recreational ketamine abuse has been linked to a variety of longer term cognitive impairments (35, 36), there is no clear evidence of ketamine or esketamine causing measurable sustained deleterious cognitive effects when used according to consensus recommendations or the FDA REMS protocol. The data for IN esketamine indicate no meaningful decreases in performance on the Cogstate Brief Battery in up to 4 years of treatment, with signals actually suggesting improvement on several measures in the group, totaling 2,769 cumulative patient-years (19). However, there was a signal showing a slight slowing of reaction time in patients over 65 years old.

In ELEKT-D (9), patients in the ECT group had a greater decline in memory performance on all measures at the end of treatment, compared with those in the ketamine group. Global Self-Evaluation scores of memory remained low through the end of the follow-up period, whereas Squire Memory Complaint Questionnaire scores gradually recovered, and the Hopkins Verbal Learning Test–Revised scores were similar across both groups at the 1-month follow up.

Cardiovascular and cerebrovascular effects.

Ketamine has been used for over 50 years as an anesthetic, and there is a wealth of information characterizing its effects on the cardiovascular system. However, most of these data have been collected in patients with serious medical illnesses necessitating surgery or other procedures requiring anesthesia, sedation, and administration at much higher doses. Because ketamine itself does not have an FDA indication for the treatment of depression, there is no coordinated registry to collect real-world data on complications arising from its use in the treatment of psychiatric disorders. Studies consistently show that ketamine can transiently increase systolic and diastolic blood pressure beginning shortly after the infusion, peaking at around 40 minutes, and typically returning to baseline within 2 hours of administration. These effects on blood pressure are characteristically greater in older and hypertensive patients (37–39).

Since the approval of IN esketamine for the treatment of depression by the FDA, there has been a tremendous amount of data collected as part of the imposed REMS, including safety data on over 34,000 patients receiving over 840,000 treatments with IN esketamine. The data show that 6.5% of patients who are treated with either 56 mg or 84 mg IN esketamine will experience significant increases in blood pressure (>185/105) during the 2 hours after administration for the first eight treatments (40). Overall, less than one in 1,000 treatments resulted in any blood pressure changes that were considered serious AEs. However, it is important to note that these data were collected in patients who were screened under the REMS that clearly aims to exclude patients with conditions that predispose them to complications from increased blood pressure, such as uncontrolled hypertension, myocardial infarction, aortic dissection, and arteriovenous malformation (41). History of intracerebral hemorrhage or aneurysmal vascular disease are absolute contraindications for IN esketamine.

Blood pressure should be controlled before starting ketamine or esketamine treatment and monitored after administration until blood pressure values return to baseline. In conditions in which an increase in blood pressure risks complications, ketamine and esketamine should be avoided, if possible.

Cardiac adverse effects of ECT have decreased with the implementation of continuous heart rate, blood pressure, and electrocardiogram monitoring but occur in as many as 7.5% of healthy patients and 55% of patients with a preexisting cardiac illness (42). Cardiopulmonary adverse effects may include hypertension, transient cardiac arrythmias, respiratory distress, aspiration, electrocardiogram abnormalities, and transient asystole (43). Factors that are known to moderate cardiac adverse effects include preexisting cardiac disease, electrode placement, anesthetic agents, subconvulsive stimulation, age, and psychotropic drugs (43).

Clinicians must assess and manage comorbid cardiovascular conditions such as hypertension, coronary artery disease, atrial fibrillation or other underlying arrythmias, and valvular diseases such as aortic stenosis before initiating ECT. It is imperative to consult with the treating cardiologist to optimize comorbid conditions and with the anesthesiologist to confirm that diagnostic information is available and to discuss management of these conditions during treatment to minimize risk. Continuous electrocardiographic monitoring is necessary for all patients, especially those with an implantable cardioverter-defibrillator, and resuscitative equipment should be readily available in case of emergencies.

Previously, brain tumors and space-occupying lesions were considered absolute contraindications to ECT because of concerns regarding elevation of intracranial pressure. However, recent evidence suggests that ECT may be safe in patients with cerebral aneurysms (44), although extreme caution is warranted.

Depression with psychotic features.

ECT has been shown to be particularly effective in major depressive disorder with psychotic features. Psychotic features are a predictor of ECT response, in part because patients with major depressive disorder with psychotic features often receive fewer antidepressant or antipsychotic trials before ECT, making them less medication resistant than patients with nonpsychotic major depressive disorder (45). Indeed, the predictive effect of psychotic features is stronger in patients with lower levels of medication resistance (46), and patients with known medication resistance are less likely to respond to ECT than those who are not medication resistant (47). It remains unclear how much the interaction between the level of treatment resistance and treatment outcome could have explained the findings in the head-to-head studies that suggested that psychotic features may have been a marker for better ECT response.

Patients with psychotic symptoms are typically excluded from ketamine and esketamine trials because of the theoretical risk of exacerbating psychosis; consequently, the strongest data for ketamine and esketamine are for the treatment of major depressive disorder without psychotic features. That said, the KetECT trial included patients with psychotic features and found that 50% of patients with depression with psychotic features remitted after treatment with ketamine with no indication of adverse reactions. A post hoc analysis of 69 patients with a history of psychosis who received ketamine or placebo found a significant improvement in depressive symptoms in those who received ketamine (48). Patients with a history of psychosis experienced more short-term dissociative symptoms than those without a history of psychosis, but effects resolved post infusion. A meta-analysis of pilot studies and case reports of patients with a history of psychosis or current psychotic symptoms treated with ketamine concluded that ketamine was both safe and effective in this population (49). Thus, the evidence cited previously does not suggest that ketamine exacerbates psychosis beyond its transient effects. However, further studies are clearly needed before the treatment can be assumed to be safe or effective in this population.

Substance use disorder.

There is little information to suggest that ECT could increase risk of substance use disorder. In comparison, ketamine is used recreationally, and ketamine misuse, dependence, and diversion are concerns. Indeed, most RCTs that investigated the use of ketamine for TRD have excluded individuals with substance use disorder. Patients should be screened for substance use disorder and closely observed during their treatment series (50, 51). The number of treatments should be limited to the minimum needed to achieve a clinical response, and, crucially, administered under medical supervision in a clinical setting because of theoretical concerns of increasing the risk of or reigniting substance use.

In recent years, ketamine use has become more widespread in the United States, both medically and recreationally. Although still rare, there has been an increase in ketamine poisonings when co-used with opioids or gamma-hydroxybutyrate (52). This underscores the importance of close monitoring of ketamine use in a controlled clinical setting (52, 53).

Age considerations.

Some studies suggest that older age is a predictor of better ECT response (46). The Prolonging Remission in Depressed Elderly, or PRIDE, study demonstrated robust effectiveness of right unilateral, ultrabrief, pulse ECT combined with venlafaxine in 240 participants ages 60 years and older with unipolar depression (54). Sixty-one percent met the criteria for remission, and 70% met the criteria for response. Odds of remission were 1.89 times greater in participants ages 70 years and older compared with participants ages 60–69 years. ECT was well tolerated; 16 serious AEs were reported in 13 patients and included disorientation, tardive seizure, atrial fibrillation, and urinary retention.

Other studies have not found an influence of age on ECT treatment efficacy. A study that compared younger (17–44), middle-aged (45–63), and older (>65) patients with TRD who received ECT found similar rates of remission across age groups (55). However, older patients had higher response rates and required fewer treatments to achieve response or remission, although the authors attributed this to other clinical differences, such as shorter duration of major depressive episode and lower levels of treatment resistance.

Most of the high-quality data for ketamine in older adults are associated with esketamine. A phase 3 trial included 138 participants, ages 65 years and older, with TRD who were treated with flexible doses of esketamine plus an oral antidepressant or an oral antidepressant and placebo. The results in symptom improvement did not demonstrate a statistical significance from that achieved with placebo alongside a new oral antidepressant drug, but the numerical improvement in scores with esketamine was within range to be deemed clinically meaningful (56). However, a subgroup analysis by age revealed a greater reduction in MADRS scores in patients ages 65–74, but not in patients ages 75 years and older, suggesting that the oldest of the population did not benefit as well from the treatment, although this subgroup was small (N=21). Treatment-emergent AEs were consistent with those seen in younger patients. A total of 64.5% of participants in the esketamine arm received the highest dose of 84 mg and tolerated it well. No new safety concerns were identified.

A post hoc analysis comparing outcomes between younger (18–64; N=624) and older (≥65; N=178) patients in an open-label, long-term study showed that a significantly greater proportion of younger than older patients were responders at day 28, which suggests that the 4-week induction period may be too short for optimal effects in the older population, especially when starting with 28 mg to ensure tolerability (57). This slower onset of treatment response may have contributed to the nonsignificant results in the phase 3 RCT. The analysis also demonstrated that rates and types of treatment-emergent AEs in both younger and older groups were similar.

Two small studies of IV ketamine in older adults, a pilot RCT and an open-label pilot study, have shown promising outcomes. The RCT of 16 participants who received subcutaneous doses (0.1–0.5 mg/kg), with 0.01 mg/kg midazolam inserted within the first three administrations, showed 68.8% remission at least at one timepoint, and 50% maintained remission for 7 days posttreatment (58). In the open-label phase, repeated treatments resulted in a longer time to relapse and improved remission rates. Adverse effects were typically mild and transient, including perceptual disturbance, minimal increases in systolic and diastolic blood pressure, and mild neurologic symptoms. The open-label study with 25 participants receiving IV ketamine at 0.5 mg/kg/40 min found 48% response rates and 24% remission rates in the acute phase, with 47% response rates and 27% remission rates in the continuation phase (59).

Subgroup analyses of the two noninferiority studies revealed contradictory findings about age and efficacy. KetECT observed a significant interaction between treatment and age group. Participants older than 50 years who received ECT had significantly higher remission rates compared with patients younger than 50 years (77% vs. 50%), whereas participants older than 50 years who received ketamine had significantly lower remission rates compared with those in the younger group (37% vs. 61%). However, ELEKT-D’s age subgroup analysis (of participants ages ≤47 years old or >47 years [median=47 years]) did not find a significant interaction with the treatment group in predicting treatment response, but this may have been limited by the smaller number of older patients participating in the study.

In summary, ketamine data in older adults are limited. Esketamine has demonstrated clinically meaningful improvement, but the time to response may be slower than ECT. Sufficiently powered studies of ECT and ketamine in older adults are needed to improve our understanding of how the two modalities compare in these subgroups.

Patient preference.

Generous amounts of positive media attention have been paid to ketamine and esketamine over the past several years, which stands in contrast to the negative views of ECT that are commonly portrayed in the media (60). These views, along with differences in the treatment’s purported mechanisms of action and AE profiles, are likely to affect patient preference.

KetECT reported higher dropout rates in the ketamine group after initiating treatment. The authors attributed this to difficulty in coping with ketamine’s dissociative side effects and knowing that ECT was available after the study. Conversely, in ELEKT-D, a larger percentage of patients randomized to ECT withdrew before treatment initiation. Thirty-one patients assigned ECT withdrew before their first treatment, compared with four patients assigned ketamine. The authors of ELEKT-D cited patient preference and logistical issues as primary reasons.

Explaining the potential side effects of both treatments and allowing the patient to weigh the comparative risks is a productive way to reconcile the distinct side-effect profiles of ECT and ketamine. Not only is including patients in the clinical decision a core tenet of ethical patient care, but shared decision making could also provide increased adherence; treatment satisfaction; and, possibly, even treatment outcome (61).

Accessibility and costs.

The treatment decision process is meaningless unless the treatment is accessible. ECT utilization has been reduced significantly in recent decades (6, 62), showing notable geographic variability, with particularly low utilization rates in the western region of the United States (6). Additionally, ECT is traditionally used in the inpatient setting, although there is a growing trend toward outpatient use (63). ECT is also less accessible to patients from minority groups (64, 65), as well as patients who are uninsured or underinsured. The cost of ECT can vary widely, with a typical course of 10 ECT treatments ranging from $10,000 to $15,000 (66). ECT is widely covered by most insurance plans. However, uninsured patients face high treatment costs.

There is a growing use of off-label ketamine for the treatment of psychiatric disorders across the country (67). Off-label ketamine generally incurs out-of-pocket expenses ranging from $2,000 to $4,000 for 2–3 weeks of treatment (68). Esketamine prescriptions rose by 121% from 2019 to 2020 after its FDA approval, with significant variability in prescriptions by state in 2020 (69). The FDA approval of IN esketamine has facilitated more insurance coverage, resulting in out-of-pocket expenses of approximately $480–$720 for 8 weeks of treatment (70). However uninsured patients face significantly higher expenses, with costs ranging from $10,800 to $15,600 (70).

Accessibility by geographic region and inpatient or outpatient setting, as well as insurance coverage, undoubtedly influence the clinical decision between treatments. Both treatments can also place a burden on the family or caregiver. In addition to financial concerns, there are logistical challenges, such as transportation to and from appointments. Ensuring accessibility to interventions for depression is critical.

Miscellaneous factors.

ECT requires the administration of an anesthetic, and the most suitable anesthetic for each patient is an important consideration. As an anesthetic agent, ketamine has a higher risk of post-ECT nausea than propofol, for example (43). For patients with cardiac disease, propofol or methohexital may be preferable (43). Other common side effects of ECT related to general anesthesia include headache, jaw soreness, myalgias, post-procedure nausea and vomiting, and fatigue, which are usually mild and transient.

Adverse effects such as prolonged apnea or prolonged paralysis, caused by the muscle relaxants given during ECT, and seizure-related adverse effects such as prolonged seizures and tardive seizures are more serious but rare.

Although ECT can be safely administered to patients with various respiratory conditions, in patients with severe respiratory disorders or respiratory failure, care must be taken with ECT, as these conditions could complicate the management of anesthesia and maintenance of a patent airway during ECT sessions (71). In rare cases, ketamine and esketamine are associated with respiratory depression (72).

Given that ketamine’s metabolism and clearance involve hepatic pathways, patients with impaired liver function may be at increased risk for adverse effects (27). Recently, the FDA issued a warning about liver injury associated with ketamine (73). ECT may be a safer modality in comorbid hepatic conditions, but the choice of anesthetic is important.

Chronic recreational ketamine use is associated with increased severe lower urinary tract dysfunction symptoms (74). Long-term safety data on esketamine suggest an increased risk of symptoms such as dysuria or urgency but no severe pathology among patients receiving the treatments clinically in accordance with the prescribing guidelines (75). Still, clinicians may want to weigh the risks and benefits of initiating ketamine treatment in patients with severe kidney dysfunction or preexisting interstitial cystitis (76).

Because of the muscle contractions induced during ECT, severe osteoporosis could be a risk factor for fractures during treatment, and ketamine might be a safer treatment if adequate paralysis cannot be achieved (77).

In summary, the treatment of depression with either ketamine or ECT requires consideration of the patient’s comorbid conditions and overall health status to optimize treatment efficacy while minimizing risks and adverse effects.

Future Work

Future research that compares ECT with ketamine and esketamine should address several unanswered questions to optimize treatment decisions for an individual patient. Studies should investigate the short- and long-term efficacy of both treatments in different settings, (e.g., inpatient vs. outpatient), across age groups, and across depression subtypes and comorbid medical conditions. An important endeavor in psychiatry is the implementation of personalized medicine, or precision psychiatry, whereby treatment approaches are tailored to an individual on the basis of a combination of clinical, biological, genetic, and lifestyle factors. Identifying biomarkers that help to inform clinicians on which patients would respond best (in terms of both efficacy and side effects) to ketamine versus ECT would be an invaluable advance in treatment planning. Although the field is not there yet in terms of predictive biomarkers, initial work points to some biomarkers of treatment response to ECT (e.g., hippocampal volume) (78) and ketamine (e.g., neurotrophic factors, inflammatory markers) (79, 80). With the adoption of big-data approaches and artificial intelligence, the use of biomarkers for optimizing treatment selection in psychiatry appears to be on the horizon, holding promise for reducing trial-and-error approaches in psychiatric treatment and improving patient care (81). Last, increasing evidence suggests that augmenting ECT and ketamine with psychotherapy, such as cognitive-behavioral therapy, could help to sustain the antidepressant benefits and extend time to relapse (82, 83). Therefore, research into the optimal ways of combining ECT or ketamine treatment with evidence-based psychotherapy is warranted.

Conclusions

In conclusion, both ECT and ketamine have demonstrated effectiveness in treating adults with moderate to severe depression. However, head-to-head comparisons have revealed variable results, and the determination of which treatment is more effective requires further clinical investigation. It is important to note that the efficacy and safety of ECT and ketamine can vary on the basis of individual patient factors. Therefore, the pertinent question is: Which patient-specific factors make a treatment more suitable for an individual? In light of the present evidence, both ketamine-esketamine and ECT should be considered for the treatment of TRD in adults. Treatment selection should be personalized on the basis of associated comorbid conditions and risks, as well as other important patient-specific factors, such as treatment accessibility and patient preference.