Here is a scenario that will show up in your clinical rotations, and probably on a practice question or two. A patient with severe, treatment-resistant depression finally gets a rapid response to ketamine or esketamine. Two weeks ago they could barely get out of bed or talk about the future. Today they are laughing in the waiting room. Somebody says, quietly, “Great, so we can wrap this up now, right?”
That moment is where good prescribers separate themselves. In the October 2026 issue of The Carlat Psychiatry Report, Benjamin D. Brody, MD, wrote “An Update on Ketamine,” and the most useful thing in it isn’t a new mechanism or a new molecule. It’s a set of answers to the questions that come after the first good response: How long do we keep going? What happens if we stop? And what do we do with the patients who don’t fit the neat trial population? If you’re training to become a PMHNP, those are exactly the questions you’ll be handed in practice, usually on a Tuesday afternoon with a full schedule.
The First Principle: Ketamine Is Preferably a Short-Term Tool
Dr. Brody starts with a framing that I wish more students heard early. Ketamine and esketamine, he writes, are preferably short-term treatments. For a patient who has had a handful of well-separated depressive episodes, the job of the drug is to help them climb out of the current episode. Then other therapies carry the long game.
The patient who needs a different plan looks different on paper. Brody describes people with a chronic course, a longer list of failed medication trials, and shorter stretches of wellness between episodes. Those are the patients who may need ketamine as an ongoing maintenance treatment. He isn’t giving you a rule about weeks or months. He’s giving you a way to think about the illness pattern, which is the habit boards test when a vignette asks for the next best step.
What the Evidence for Continuing Actually Shows
Most of the long-term data we have comes from esketamine (Spravato), the nasal spray, rather than from IV racemic ketamine. Brody points to the SUSTAIN program, a set of trials that followed patients for up to four and a half years and turned up no new safety problems with continued treatment.
The study that matters most for the “should we stop?” question is SUSTAIN-1, published by Daly and colleagues in JAMA Psychiatry in 2019. Here is how it worked. Patients first received 16 weeks of esketamine plus a newly started oral antidepressant. Those who reached stable remission or stable response were then randomized to either keep the esketamine or switch to a placebo nasal spray, with the oral antidepressant continued in both groups. In total, 297 people were randomized.
Among stable remitters, 26.7 percent of those who stayed on esketamine relapsed, compared with 45.3 percent of those switched to placebo. That works out to a 51 percent reduction in relapse risk. Among stable responders who hadn’t fully remitted, relapse was 25.8 percent on esketamine versus 57.6 percent on placebo, a 70 percent reduction. Carlat summarizes this as a 50 to 70 percent reduction in relapse risk, and notes the benefit was larger for patients who had not fully recovered.
Pause on that. The people with the most to lose from stopping were the ones who looked a little less polished.
One Head-to-Head Trial, and Why You Should Read the Fine Print
Brody also cites Reif and colleagues’ 2023 New England Journal of Medicine trial comparing esketamine with extended-release quetiapine as augmentation in treatment-resistant depression. In this eight-month comparison, short-term remission was 27 percent for esketamine versus 18 percent for quetiapine, and over the long run the esketamine patients were 1.6 times more likely to stay euthymic.
Now here is where I want you to develop a skill that no flashcard deck will give you: reading a result with one eyebrow raised. When The Carlat Psychiatry Report reviewed this trial in 2023, the reviewer, Ivy Song, MD, pointed out that the difference on the depression rating scale was only about 2.8 points out of 60 at week 8, which raises a fair question about clinical meaning. She also flagged that the manufacturer of esketamine funded the study, and that none of the esketamine patients had tried the drug before, while some quetiapine patients had likely already failed other atypical antipsychotics. She still considered the comparison valuable.
That is the right posture. You don’t throw the finding out, and you don’t swallow it whole. You note who paid, who was in the comparison arm, and how big the effect was in real units.
The Withdrawal Question Everyone Asks
Students (and plenty of seasoned clinicians) worry about whether you can become dependent on ketamine and go through withdrawal when you stop. Brody’s answer is measured. A withdrawal syndrome has been reported in the literature, but he has never seen one in his own practice, and he doesn’t expect it at therapeutic doses and frequencies. Some of his patients stay on ketamine once or twice a month long term, and he considers that probably safe at usual therapeutic doses. If a taper is considered, it is usually slow, with treatments spaced further and further apart.
The bigger risk is the one that gets less airtime: relapse. Think of it this way. The danger of stopping isn’t that the patient will feel awful from the drug leaving the body. It’s that the depression returns, sometimes quickly, in a patient who may have been suicidal not long ago. Brody’s warning is blunt. Don’t give ketamine to a suicidal patient, watch them improve quickly, and then stop abruptly. The effects are often transient, typically lasting a few weeks, so the plan for what comes next needs to be in place before the first infusion, not after the last one.
For exam purposes, hold onto the contrast. A drug’s discontinuation risk can be pharmacologic (a withdrawal or discontinuation syndrome) or it can be about the illness (relapse). With ketamine, the clinically important risk after stopping is mostly the second kind. Test writers love that distinction, and so do real patients’ families when they ask why you’re being cautious about ending a treatment that seems to be working.
The Patient the Trials Didn’t Picture: Borderline Personality Disorder
This section of Brody’s update is short, and it’s the one I’d tell you to remember if you work in community mental health, where borderline personality disorder is common and where these patients are often treated for depression on top of everything else.
The research is limited, but here is what he summarizes. A 2025 paper by Scott and colleagues in the Australian and New Zealand Journal of Psychiatry suggests patients with borderline personality disorder respond similarly for depression but may carry a higher risk of dissociation. Another 2025 study by Hu and colleagues in Psychiatry Research suggests ketamine may increase empathy and improve mentalization, which could support therapeutic engagement. And Brody’s bottom line is a sentence worth writing on a sticky note: ketamine may augment psychotherapies like dialectical behavior therapy, but it does not replace them.
Look at how that reframes your role. As a PMHNP you are not merely the person who writes the order. You are the person who makes sure the drug is serving the treatment plan rather than substituting for it. A patient who is more dissociative on treatment needs closer monitoring and a conversation about what they’re experiencing, not a shrug and a rescheduled appointment.
The Negative Trial That Got Everyone Arguing
Every field needs a result that makes people uncomfortable, and in this issue it’s the Irish trial by Jelovac and colleagues, published in JAMA Psychiatry in 2025. It was a pragmatic, double-blind trial of IV ketamine versus midazolam as an active control in 65 hospitalized patients, and the ketamine did not statistically outperform the control. According to a separate Carlat review of the same study by Victoria Hendrick, MD, patients received up to eight twice-weekly infusions alongside usual inpatient care. Response was 47 percent with ketamine versus 33 percent with midazolam, and remission was 44 percent versus 30 percent. Those are bigger numbers on the ketamine side, but the difference didn’t reach statistical significance.
So is ketamine overrated? Brody, to his credit, doesn’t jump to that conclusion, and neither should you. He praises the effectiveness-research approach, then names the problems. Only non-suicidal patients were enrolled, which limits how well the findings apply to American inpatients, most of whom are suicidal. With roughly 30-some patients per arm, the study may have been underpowered. And many participants correctly guessed which treatment they received, which is a classic problem with ketamine trials, since the dissociative effects are hard to hide. Dr. Hendrick adds that midazolam is not an inert placebo and can have short-term mood effects of its own, which may mask a real ketamine signal.
If you take one research-literacy lesson from this section, make it this: a negative trial is not the same as a trial that disproves the treatment. Ask who was enrolled, how many, how well blinding held up, and whether the comparison arm was truly inactive. Then ask what question the trial actually answered.
Putting It to Work on Your Rotations and Your Boards
Let’s turn all of this into things you can actually do.
Start with the illness pattern, not the drug. Before you decide how long ketamine or esketamine should continue, ask how many episodes, how many failed trials, and how long the patient stays well between episodes. Brody’s short-term-versus-maintenance split follows directly from those answers.
Plan the exit before the entrance. What will keep this patient well when the infusions or nasal spray sessions taper? The oral antidepressant, psychotherapy, a follow-up schedule, a safety plan. In SUSTAIN-1 even the esketamine group kept taking an oral antidepressant. Ketamine wasn’t a replacement for the rest of the treatment. It was an addition to it.
Watch the first weeks after stopping. If the effect often lasts only a few weeks, that’s your highest-risk window, especially for a patient who started treatment with suicidal thoughts.
Know your evidence by name. SUSTAIN-1 (Daly, JAMA Psychiatry 2019) for relapse prevention, Reif (NEJM 2023) for esketamine versus quetiapine, and Jelovac (JAMA Psychiatry 2025) as the cautionary inpatient trial. Being able to say which study supports which claim will serve you in a case conference as much as on an exam.
Remember what you don’t know. The long-term data are mostly esketamine data. Brody himself calls his comfort with twice-monthly ketamine a judgment that it is “probably safe,” not a guarantee. It’s okay to hold that kind of uncertainty out loud with a patient. In fact, it builds trust.
The Bigger Lesson
The relapse data are impressive. The head-to-head trial has real limitations. So does the negative trial. A prescriber who can hold all of that at once and still make a calm, individualized plan is exactly who this field needs.
And if you’re sitting there tired, juggling clinical hours and a study schedule, thinking that you’ll never keep all this straight, I want you to hear something. You don’t need to memorize every percentage in this post. You need to understand the shape of the story: rapid response, high relapse risk when you stop, a maintenance decision based on the illness pattern, and a plan that includes more than the drug. Once you have the shape, the details have somewhere to land.
When you’re ready to turn knowledge like this into a passing board score, we built PMHNPReview.org to help you do exactly that, with practice questions and clear explanations designed around how the exam really asks about these decisions. Come take a look, and keep going. The patients who will walk into your office in a year are counting on the person you’re becoming right now.
Sources discussed: Benjamin D. Brody, MD, “An Update on Ketamine,” The Carlat Psychiatry Report, October 2026; Daly EJ et al., JAMA Psychiatry 2019;76(9):893–903 (SUSTAIN-1); Reif A et al., N Engl J Med 2023;389(14):1298–1309, as reviewed in The Carlat Psychiatry Report; Jelovac A et al., JAMA Psychiatry 2025;82(12):1216–1224, as reviewed in The Carlat Hospital Psychiatry Report; Scott E et al., Aust N Z J Psychiatry 2025; Hu H et al., Psychiatry Res 2025. This post is educational and is not a substitute for clinical judgment or supervision.