When a new trial doesn’t rewrite the 30-year story of ketamine

As the co-founder of Ember Health and a member of American Society of Ketamine Physicians, Psychotherapists, and Practitioners (ASKP³), my job is to stay up to date on the latest research in ketamine for depression. That’s because at Ember Health, we believe in pairing compassion with evidenced based, data-driven care.

When the headlines (such as “Clinical Trial Finds Ketamine Not Effective For Depression” – U.S. News) around the KARMA-Dep 2 study in JAMA Psychiatry, led by Trinity College in Dublin, suggested that intravenous ketamine may not help depression, I knew it was necessary to take a closer look. Here, I aim to reflect on the findings publicly to invite a healthy discourse.

In short, this is an important paper, but it requires context to interpret the results, and does not negate the wealth of evidence surrounding IV ketamine as an effective treatment for depression.

The trial, called KARMA-Dep 2, compared two groups of patients hospitalized for depression: the first received repeated ketamine infusions, and the second received repeated doses of midazolam.

Midazolam is a short-acting benzodiazepine medication commonly used for its sedative, anxiolytic (anti-anxiety), muscle relaxant, and amnestic (memory-impairing) effects. Midazolam itself is not an antidepressant and is not approved or typically used to treat depression directly. However, it sometimes appears in depression research or clinical discussions because of its role as a control or comparator drug in studies testing ketamine and other rapid-acting antidepressants.

Intravenous (IV) ketamine, on the other hand, was originally developed as an anesthetic and began to be used and studied as an exciting new antidepressant treatment 30 years ago. Today, it’s been investigated in over 300 studies and clinical trials as a safe, effective treatment for (treatment-resistant) depression when used off-label. It has in fact been the most studied drug to treat depression in the last 30 years. It has around double the efficacy of SSRIs at alleviating the symptoms of depression, with extremely minimal side effects and high bioavailability, allowing for low, sub-anesthetic doses. The FDA has recommended against any route of ketamine administration that is not IV.  

In the study, both groups improved significantly, but the difference between them was not statistically significant. Here are five core reasons why I believe the study’s results should be interpreted with care:

1. Patient selection

Out of about 3,000 people screened, only 62 were enrolled. That very low selection rate raises questions about representativeness, particularly as all suicidal patients where screened out, which is a group that is highly responsive to ketamine. The chosen sample focused on people who were hospitalized for nearly 4 weeks (which simply doesn’t happen in the U.S., particularly not for patients who are not suicidal), and does not remotely reflect the broader outpatient population where most ketamine therapy occurs and where most studies have shown success.

2. Inpatient setting and level of care

This was a study of inpatient care, not outpatient care. It was also a study in a well resourced healthcare system. A four-week psychiatric inpatient admission is exceedingly rare in the U.S., representing a very specific care environment. Inpatient cases in general are more acute, on more medications, and receive far more intensive support, including daily provider visits and 24-hour nursing care. Those alone would be expected to drive substantial improvement across both study arms. So this study didn’t demonstrate a “placebo response”, it demonstrated a “response to high quality inpatient care”. That effect is real, and can create a ceiling that makes it difficult to see incremental benefit from any medication.

3. Polypharmacy and comparator choice

Most participants in the ketamine arm were on benzodiazepines and antipsychotics, both known to blunt ketamine’s antidepressant effect. Midazolam, the comparator, is itself a benzodiazepine with and has significant short term antidepressant effect. Equally challenging, over 90% of the people in this study correctly identified their treatment arm, which means there was no actual blinding accomplished in this trial. Comparing ketamine against another active, mood-elevating drug in a non-blinded way makes detecting meaningful differences that much harder.

4. Timing of assessment

The trial evaluated outcomes well after the last infusion. Ketamine’s antidepressant effects tend to appear rapidly in days, and often wanes over weeks. Assessing too late (6 months later) misses the window when ketamine’s impact is strongest, making the results appear smaller than they truly are.

5. Outcome interpretation

When you look at the actual data from the study, ketamine did improve symptoms more than placebo. The differences did not reach statistical significance. Equally important, both groups in the study did better from their care, by about 50%, which is consistent with what we see in other inpatient antidepressant trials.

In summary

This trial adds to our understanding, but does not change the broader evidence base. It tells us that in a highly resourced, multi week hospital environment, with high benzo use, ketamine did not incrementally improve outcomes. More than 50 randomized controlled trials and over 140 clinical studies still show that ketamine, when used thoughtfully and in the right setting, can be a transformative treatment for depression.

Gratefully,
Dr. Nico Grundmann
Co-Founder, Ember Health