Depression relieved within hours, not weeks

Researchers at Yale University gave seven depressed patients a single low-dose infusion of ketamine in 2000 and measured a drop in depression scores within three days. Six years later a second laboratory saw the change within two hours.

A drug built to put people to sleep

Ketamine was never meant to be an antidepressant. It was developed in the 1960s as an anesthetic, a compound that could carry a patient through surgery while keeping breathing and blood pressure steady. Surgeons used it. Veterinarians used it. By the 1990s it had also acquired a second life on the street, where its ability to detach the mind from the body made it a club drug.

That detachment is what first drew psychiatrists to it, and not because they thought it would help anyone. In 1994, John Krystal and colleagues at Yale gave nineteen healthy volunteers a forty-minute infusion of ketamine at two low doses, 0.1 and 0.5 milligrams per kilogram, under randomized, double-blind, placebo-controlled conditions. The point was to build a temporary model of psychosis. The drug produced behaviours resembling the positive and negative symptoms of schizophrenia, altered perception and impaired tests of memory and attention, all of which faded as the infusion wore off. Ketamine was a tool for studying what goes wrong in the brain. Nobody was looking for what might go right.

Seven patients, two days, one number

The question changed because of a theory. Through the 1990s a growing body of animal work suggested that glutamate, the brain's main excitatory messenger, might be involved in depression and in how antidepressants work. Ketamine blocks one glutamate receptor, the N-methyl-D-aspartate receptor, usually shortened to NMDA. If glutamate mattered, a drug that blocked NMDA receptors might do something to mood.

In 2000, Robert Berman and colleagues at Yale, with Krystal among the authors, published a four-page paper in Biological Psychiatry. Seven patients with major depression completed two test days. On one day they received an intravenous infusion of ketamine at 0.5 milligrams per kilogram; on the other, saline. The order was randomized and neither the patients nor the clinicians rating them knew which was which. It was the first placebo-controlled, double-blind trial of a single dose of an NMDA antagonist in depression.

The result was measured on the 25-item Hamilton Depression Rating Scale, a clinician's checklist in which higher scores mean deeper depression. Within 72 hours of the ketamine infusion, scores fell by a mean of 14 points, with a standard deviation of 10. After the saline infusion, they fell by a mean of zero. Every antidepressant then in use took weeks to separate from placebo. Here was separation inside three days, from one dose.

Seven patients is a very small study, and the authors said so. They wrote only that the results "suggest a potential role" for drugs that modulate the NMDA receptor. The paper did not announce a new treatment. It announced a new question.

Why it took six years to be believed

A single small trial from one laboratory does not change a field, and this one had reasons to be doubted. Ketamine was a known drug of abuse. Its dissociative effects made a true blind difficult, because patients can often tell when they have received the active drug. And the idea that depression could lift in hours contradicted a decades-old assumption that antidepressants work slowly because the brain must slowly adapt to them.

The replication that mattered came in 2006 from the National Institute of Mental Health. Carlos Zarate and colleagues enrolled eighteen patients whose depression had failed to respond to earlier treatment. After two weeks off medication, each received ketamine at 0.5 milligrams per kilogram on one day and placebo on another, a week apart, in a randomized, double-blind crossover. Ratings were taken at 40, 80, 110 and 230 minutes and then at one, two, three and seven days.

Depression scores on ketamine separated from placebo within 110 minutes of the infusion and stayed separated for a week. The effect size at 24 hours was 1.46, which in clinical trials counts as very large; most antidepressant trials report effect sizes a fraction of that after six weeks. Of the seventeen patients who received ketamine, 71 percent met criteria for response and 29 percent for remission the day after the infusion. About a third held their response for at least a week.

That is where the word "hours" in this story's title comes from. Berman had shown improvement inside three days. Zarate showed it inside two hours.

What was happening in the brain

The speed demanded an explanation, and in 2010 one arrived from the laboratory of Ronald Duman at Yale. Nanxin Li and colleagues reported in Science that in rats, a single dose of ketamine rapidly switched on a signalling pathway called mTOR in the prefrontal cortex, increased the production of proteins used to build synapses, and increased the number and function of new spine synapses, the small protrusions where one neuron receives input from another.

Two details made the finding more than a correlation. When the researchers blocked mTOR, ketamine no longer built synapses and no longer produced antidepressant-like behaviour in the animals. And the changes ran in the opposite direction to what chronic stress does to the same brain region, where spines are lost. The proposal was simple: depression involves a loss of connections, ketamine rebuilds them fast, and that is why the mood change does not wait.

This was rat work, and it does not prove the same thing happens in a human brain. But it gave the clinical finding a mechanism to be tested against, and it reframed the drug. Ketamine was no longer a chemical that merely dulled a receptor. It was a chemical that appeared to trigger repair.

The larger trial

By 2013 the field needed a trial big enough and controlled well enough to settle whether ketamine's effect was real or an artefact of patients knowing what they had received. James Murrough and colleagues ran it at two sites with 73 patients with treatment-resistant depression. Instead of saline, the comparison was midazolam, a sedative that produces noticeable effects of its own and therefore a fairer blind.

Twenty-four hours after a single infusion, depression scores on the Montgomery-Åsberg scale were 7.95 points lower in the ketamine group than in the midazolam group. Sixty-four percent of ketamine patients responded, against 28 percent on midazolam. The authors were careful to say that durability and safety needed more work before clinical use. But the effect had survived an active placebo, a larger sample and a second site.

From infusion to nasal spray

Ketamine is an old, cheap, generic drug, which is one reason no company rushed to run the trials that would turn it into an approved antidepressant. The path to a product ran instead through esketamine, one of the two mirror-image molecules that make up ketamine, delivered as a nasal spray. On March 5, 2019, the U.S. Food and Drug Administration approved esketamine, sold as Spravato, for adults with treatment-resistant depression, to be used alongside an oral antidepressant.

The approval came with conditions that reflect the drug's history. Because of the risk of sedation, dissociation and misuse, esketamine can only be given in a certified doctor's office or clinic, where the patient is observed afterwards, and it is distributed under a Risk Evaluation and Mitigation Strategy. It was the first antidepressant with a new mechanism of action to be approved in decades.

What the story does and does not show

The 2000 paper showed a difference in seven people over three days. The 2006 paper showed it in eighteen people within two hours, lasting a week in a third of them. The 2013 paper showed it in 73 people against an active placebo at one day. None of these studies showed that a single infusion cures depression; the effect fades over days to weeks, and how to sustain it safely is still being worked out. The mechanism was established in rats, not people.

What the sequence does show is how a finding earns its place. A small trial posed a question the field was not ready for. A second laboratory reproduced it with tighter timing. A third explained how it might work. A fourth tested it against a control that could fool patients. Nineteen years after seven patients at Yale, a regulator approved a descendant of the drug under strict supervision. The first rapid antidepressant was an anesthetic, and it took two decades of measurement to prove it.

Hours, not weeks. The first rapid antidepressant.