Ketamine vs electroconvulsive therapy: dissecting the surprises

Ketamine has astonishing efficacy in patients with major depressive disorder; about half of treated patients, or more, can expect to respond, and many may also remit, within just 1–2 days of receiving a single dose of the drug; in contrast, remission usually takes 2–3 weeks of alternate-day sessions with electroconvulsive therapy (ECT) or 6–8 weeks of daily dosing with antidepressant drugs. A limitation is that rapidly-elicited response and remission with ketamine is transient; in most patients, the benefits wear off in a week. The benefits can be extended by administering another dose of ketamine when the benefits wear off, or by administering ketamine as a course of treatment, much as is done with ECT.

Ketamine has much potential for use as an antidepressant if regulatory hurdles can be overcome; the opportunities for developing countries, such as India, are especially large for racemic ketamine administered by the oral route. ¹

Given the promises, and given the negligible risk of amnestic adverse effects as seen with ECT, a question that arises is, “How does ketamine compare with ECT”? This question has been addressed in randomized controlled trials (RCTs) that have been combined in meta-analysis. In summary, a systematic review and meta-analysis of 5 RCTs (pooled N=278) found that ECT was superior to ketamine with a small to medium effect size that was statistically significant after disregarding RCTs of poor methodological quality (standardized mean difference, -0.45; 95% confidence interval [CI], -0.75 to -0.14). ECT was also associated with significantly better chances of response (relative risk [RR], 1.27; 95% CI, 1.06–1.53) and remission (RR, 1.43; 95% CI, 1.12–1.82). ²

But a new RCT with surprising findings, conducted in the US, has now been published. ³ This is the largest study in the field, to-date. It was an open-label noninferiority comparison of thrice weekly right unilateral ECT at 6x seizure threshold vs twice weekly IV ketamine (0.5 mg/kg across 40 min) in 403 patients with treatment-resistant major depression, without psychosis, who had been referred for ECT. The number of ketamine and ECT sessions was based on clinical judgement but was proposed to extend across 3 weeks. Ongoing psychotropic medications were continued unchanged during the study.

The study found that ketamine was noninferior to ECT on efficacy outcomes. Unexpectedly, the numbers were (nonsignificantly) better with ketamine. As examples, the primary outcome, response on the Quick Inventory of Depressive Symptomatology Self-Report, was observed in 55% vs 41% of ketamine vs ECT patients, respectively. Remission on the Montgomery-Asberg Depression Rating Scale (MADRS) was observed in 38% vs 22% of ketamine vs ECT patients. Depression ratings improved by 15 vs 13 MADRS points in ketamine vs ECT groups. Expectedly, ECT was associated with cognitive impairment, and ketamine with dissociation.

Thus, in contrast with the results from meta-analysis, this largest RCT of ketamine vs ECT unexpectedly found that ketamine was (numerically) superior to ECT, which is considered to be a gold standard treatment for depression. So, what should the reader think?

The RCT had many unusual characteristics, few of which were acknowledged by the authors. For starters, of 403 patients who had been randomized, 4 ketamine patients and 31 ECT patients dropped out of treatment before starting treatment. This suggests that many patients assigned to ECT were unhappy with the treatment to which they had been assigned. In other words, the unblinded, modified intent-to-treat sample may have been biased in favor of ketamine from the start of the study; so, placebo responses could be expected to be higher with ketamine.

Next, only 11% of the sample comprised inpatients. This suggests that the sample was not typical of patients referred for ECT despite the mean MADRS score of 32 at baseline (ECT samples typically require inpatient care). Third, patients with psychotic symptoms were specifically excluded from recruitment. Depression with psychotic features is known to respond poorly to antidepressant drugs but responds very well to ECT.^4–6 So, again, the sample was less than typical for ECT, and thereby possibly biased against ECT.

Fourth, the average patient had a body-mass index of approximately 30, indicating obesity; so, patients would have received a higher absolute dose of ketamine. At fixed per kg body weight doses, higher absolute ketamine doses have been associated with better treatment response. ⁷ Thus, dosing could have favored ketamine. In contrast, fifth, ECT was suboptimally administered. Unilateral electrode placement was employed with option to later switch to bilateral treatments. Severely ill patients are usually treated with bilateral ECT from the outset. Bilateral electrode placement is less sensitive to the magnitude of suprathreshold dosing ⁸ and may be the gold standard for many patients who receive ECT. ⁹ So, starting with unilateral ECT could have diminished the potential benefits of ECT.

Sixth, most patients received 6 ECTs but only 70% received 9 ECTs. Given that about 39% of patients were switched from unilateral to bilateral ECT, perhaps many patients did not receive an adequate exposure to bilateral ECT and hence an adequate course of treatment at the 3-week study endpoint. That is, the number of ECTs administered could also have been suboptimal.

Seventh, a quarter of the patients in the sample were receiving anticonvulsants and nearly a third were receiving benzodiazepines. It is not clear how these were dosed during the study with regard to proximity to the ECT session. Both classes of drugs can interfere with the ECT seizure and potentially diminish the benefits of ECT. Finally, only 41% of patients responded to ECT and only 22% remitted. These numbers are so surprisingly low that it strongly suggests that the sample was atypical for ECT and/or that ECT was suboptimally administered.

In summary, the study design and execution were loaded to favor ketamine over ECT. This was probably not deliberate; it merely reflects ECT practice and the realities of clinical trial recruitment and execution in the US. All that this study demonstrates is that iv ketamine appears noninferior to suboptimally administered ECT in a sample that is not representative of patients who usually receive ECT. Nevertheless, the good response and remission rates in this sample of patients with refractory depression who had been referred for ECT suggests that a trial of ketamine may be justified before a trial of ECT.