Anxiety that does not respond to the first medication: my next steps

About half of patients with an anxiety disorder respond to the first medication they are given. The other half end up in offices like mine, often convinced that medication does not work for them. Usually that is not what happened. Here is how I work through anxiety that has not responded to the first attempt, in the order I actually do it.

Step one: was it a real trial?

The most common reason a first anxiety medication "failed" is that it was never given a fair chance. SSRIs and SNRIs, the first-line medications for generalized anxiety disorder, panic disorder and social anxiety disorder, take longer to work for anxiety than for depression. I tell patients to expect eight to twelve weeks at a therapeutic dose before we judge the result. A prescription stopped at week three because it had not helped yet, or because of early jitteriness that would have passed, is not a failed trial. It is an interrupted one.

Dose matters as well. Anxiety disorders often need the upper half of the dosing range, and a patient left at the starting dose for six months has not had an adequate trial. When I review a prior medication, I want to know the dose, the duration at that dose, and why it stopped. Often the right next step is the same medication, done properly.

I also ask about how it was started. Patients with panic disorder in particular are sensitive to the activation that SSRIs can cause in the first two weeks. Starting at half the usual dose and moving up slowly avoids the early dropout that gets recorded as treatment failure.

Step two: is the diagnosis right?

Anxiety is a symptom of many conditions, and a medication aimed at the wrong one will not work. Before I change anything I reconsider what we are treating.

• Obsessive-compulsive disorder often presents as "anxiety" and responds to SSRIs only at higher doses and over longer periods than other anxiety disorders, with exposure-based therapy as a necessary partner.
• Post-traumatic stress disorder can look like generalized anxiety until someone asks about the trauma.
• Attention deficit disorder in adults frequently produces chronic worry about performance and deadlines that is really a downstream effect of inattention.
• Bipolar spectrum illness is the one I most want to catch, because an antidepressant without a mood stabilizer can make it worse.
• Medical causes, especially thyroid disease and cardiac arrhythmia, are cheap to check and embarrassing to miss.

I use the GAD-7, published by Robert Spitzer and colleagues in the Archives of Internal Medicine in 2006, to put a number on the anxiety itself, but the GAD-7 does not tell me what is causing it. The history does.

Step three: what is feeding it?

Caffeine, alcohol, cannabis and stimulants are the substances I find most often behind anxiety that will not settle. Four cups of coffee a day is enough to defeat a reasonable dose of sertraline. Nightly alcohol produces rebound anxiety the next afternoon that patients rarely connect to the drink. Cannabis, which many patients use to self-treat anxiety, worsens it in a substantial share of regular users. Untreated sleep apnea and chronic sleep deprivation both produce anxiety that no SSRI will fix.

I ask about all of these at every visit, without judgment, because the answers change the plan.

Step four: switch, augment or add therapy

If the trial was adequate, the diagnosis holds and the lifestyle drivers are addressed, the medication really did not work, and I have three moves.

Switch. For a patient with no response at all, I switch, usually to a different SSRI or to an SNRI such as venlafaxine or duloxetine. Roughly half of patients who do not respond to one SSRI respond to another, so a within-class switch is reasonable, and a change of class is my next choice after that.

Augment. For a patient with a partial response, I keep what is working and add to it. Buspirone is my usual first augmentation for generalized anxiety; it is well tolerated and non-sedating, though it takes several weeks. Hydroxyzine helps as an as-needed option for acute anxiety without the dependence risk of benzodiazepines. Pregabalin has good evidence for generalized anxiety in European trials and guidelines, although it is not FDA-approved for that use in the United States, and I use it selectively.

Add therapy. This is the move I make most often, and I should make it earlier than I sometimes do. Cognitive behavioral therapy has an evidence base for anxiety disorders that is at least as strong as medication. The 2008 meta-analysis by Stefan Hofmann and Jasper Smits in the Journal of Clinical Psychiatry found large effects across the anxiety disorders, and combined treatment outperforms either alone for many patients. The United Kingdom's NICE guideline for generalized anxiety disorder places CBT alongside medication at the same step. A patient who has had two medications and no therapy has not had a full course of treatment for anxiety.

What I avoid

Benzodiazepines work fast and that is exactly the problem. For panic disorder or severe anxiety in the first weeks of an SSRI, a short, time-limited course can be appropriate, and I will prescribe one with a stop date. I do not use them as a long-term solution for anxiety that has not responded to a first medication, because tolerance, dependence, cognitive effects and the fall risk in older adults make them a poor trade. The World Federation of Societies of Biological Psychiatry guidelines and most national guidelines say the same. When a patient arrives already taking one daily, I do not stop it abruptly; we build a plan and taper slowly.

I am also cautious about adding a third or fourth medication to a regimen that is not working.

When anxiety and depression travel together

Most of my anxious patients are also depressed, and the combination is harder to treat than either alone. For these patients the medication choice leans toward agents that cover both, and when two adequate trials have failed, the interventional options I oversee come into the conversation. Deep TMS was FDA-cleared in 2021 for major depression with comorbid anxiety, and in our Palm Beach Gardens program we see anxiety scores fall along with depression scores in patients who respond. TMS is not a treatment for anxiety on its own, and I do not present it as one. For anxious depression it is a reasonable step.

How I know when we are done

I track the GAD-7 and the functional goals we set at the first visit. A score under 5 is remission. Function, as always, is the real target: the patient who can fly again, give the presentation, drive on the highway, sleep through the night. Anxiety treatment that reduces the score but leaves the avoidance in place is unfinished, and that is usually where therapy earns its keep.

If your first medication did not work, the honest answer is that very little has been ruled out. It is a careful look at what was tried, what is really being treated and what else is feeding the fire, followed by one deliberate change at a time.

More from Dr. Husted

• Why I measure depression treatment by function, not just mood2026-06-11
• When to consider Deep TMS: how I decide with my patients2026-07-09
• Spravato in practice: what I tell patients before their first session2026-08-06

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