How I think about changing an antidepressant that has stopped working

One of the most common conversations in my office begins the same way: "It worked for a while, and now it doesn't." A patient who did well on an antidepressant for months or years notices the old symptoms creeping back. Sometimes the return is gradual, sometimes it follows a stressful event, and sometimes there is no obvious reason at all. The question is what to do about it, and the answer is less obvious than "switch to something else."

First, I ask whether it ever really worked

Before I change anything, I want to know what "worked" meant. Did the patient reach full remission, with normal sleep, energy and interest, or did they get to a tolerable version of depressed? The distinction matters. The STAR*D trials, the largest study of antidepressant treatment in real-world patients, found that only about a third of people reached remission on their first antidepressant, and many of the rest were left with residual symptoms. Residual symptoms predict relapse. A patient who was "better but not well" and has now slipped back is in a different situation from one who was fully well for three years and has relapsed.

I also ask about the dose and how long it has been taken at that dose. A medication that has been at a low starting dose for two years has not stopped working; it never had a full opportunity to work. Pushing the dose into the therapeutic range is often the simplest fix and the one most often skipped.

Then I check what else has changed

Antidepressants rarely stop working for no reason. Before I blame the drug, I look for what else moved. Common culprits include:

• Missed doses. Patients often take a medication less consistently once they feel well. I ask directly and without judgment, because the answer changes the plan completely.
• Alcohol. Even moderate drinking undermines antidepressants and worsens sleep. An increase in drinking during a stressful period is a frequent explanation for a "failed" medication.
• New medical problems. Hypothyroidism, anemia, vitamin B12 or vitamin D deficiency, untreated sleep apnea and chronic pain can all produce depressive symptoms. I order laboratory work when the history suggests any of them.
• New medications. Beta blockers, corticosteroids, some hormonal treatments and certain pain medications can lower mood or interact with the antidepressant.
• Life events. A loss, a job change or a caregiving burden can overwhelm a medication that was adequate under ordinary conditions. That does not necessarily mean the medication is wrong; it may mean the patient needs more support alongside it.

Finally, I revisit the diagnosis. A patient whose depression keeps returning despite treatment, who has periods of unusual energy, reduced need for sleep or impulsive decisions, may have bipolar disorder, and an antidepressant alone can make that worse. Missing this is one of the most consequential errors in outpatient psychiatry, and it is why I screen for it at every evaluation and again when a treatment stops working.

Partial response or no response

Once I have ruled out the explanations above, I sort the situation into one of two categories, because they call for different moves.

If the patient is getting some benefit from the current medication, I lean toward augmentation: keeping the antidepressant and adding a second agent. The options with the best evidence include bupropion, which also helps with energy and sexual side effects; lithium at low doses, which has decades of data and a specific effect on suicidal thinking; certain second-generation antipsychotics such as aripiprazole, which are FDA-approved as add-on treatments for depression; and, in selected cases, thyroid hormone. The American Psychiatric Association's practice guideline for major depressive disorder supports all of these strategies, and the choice among them depends on the patient's other symptoms and medical history.

If the patient is getting little or nothing from the current medication, switching makes more sense. There is no point adding a second drug to a first one that is contributing nothing except side effects.

Switching within a class or across classes

When I switch, I have to decide whether to try another medication in the same class, for example a second SSRI, or move to a different class such as an SNRI, bupropion, mirtazapine or a tricyclic. STAR*D found that switching within the class and switching across classes produced similar remission rates on average, so the choice is driven by the individual patient. If the first SSRI produced sexual side effects or emotional blunting, I move away from the class. If it was well tolerated but simply did not help enough, a different SSRI is a reasonable second try. If the patient has prominent fatigue, low motivation or weight gain, bupropion is attractive. If insomnia and poor appetite dominate, mirtazapine may fit. Pain alongside depression points toward duloxetine.

I also pay attention to what has helped a close family member. Response to a specific antidepressant runs in families more than most people realize, and it is a cheap, useful clue.

How I make the change

The mechanics of switching matter more than patients expect. There are three broad approaches:

• Direct switch: stop one medication and start the other the next day. This works for many SSRI-to-SSRI changes at moderate doses.
• Cross-taper: gradually lower the old medication while gradually raising the new one over one to several weeks. This is my default when the medications are from different classes or when the patient is sensitive to discontinuation symptoms.
• Taper, washout, then start: stop the first drug entirely and wait before beginning the next. This is mandatory when a monoamine oxidase inhibitor is involved in either direction, because the combination can cause serotonin syndrome or a hypertensive crisis. It also applies when moving to or from certain other medications with long half-lives.

Discontinuation symptoms are real and under-recognized. Stopping paroxetine or venlafaxine abruptly can cause dizziness, nausea, electric-shock sensations and irritability within days. Fluoxetine, with its very long half-life, rarely causes them. I warn patients what to expect from the specific drug they are leaving so that they do not mistake withdrawal for a worsening of their depression.

What I tell patients to expect

A new antidepressant needs four to six weeks at a therapeutic dose before I judge it, and up to eight to twelve weeks for a full response. During the first two weeks, side effects usually arrive before benefits, and I say so in advance. We use a rating scale at each visit so that we are comparing numbers rather than impressions. If there is no improvement at all by week four, I do not wait another two months hoping; I raise the dose or move on.

I also tell patients that two failed trials is the point at which the conversation changes. That is the standard definition of treatment-resistant depression, and it opens options, including Spravato (esketamine) at our Boynton Beach office and transcranial magnetic stimulation through Stedfast Psychiatry, that are not appropriate earlier but are very much appropriate then. Changing an antidepressant is not a sign that treatment is failing. It is how treatment works.

More from Dr. Gobar

• Psychosis is treatable: what I tell families at the first visit2026-07-08
• Work stress or something more: how I tell the difference2026-08-05
• Treatment-resistant depression: the order I try things in, including Spravato2026-09-02

Educational content and not a substitute for evaluation. If you are in crisis, call or text 988.