Do Antidepressants Actually Work for Anxiety?

The name is a historical accident, and it confuses people.

When a patient comes in with panic attacks or constant worry and I suggest an SSRI, a reasonable share of the time the response is some version of: but I'm not depressed. Fair. The drug class got named for the first thing it was approved to treat, and the name stuck. If we'd discovered the anxiety benefit first, we'd probably call them something else and this conversation would go differently.

So, plainly: yes. SSRIs and SNRIs are first-line medication treatment for most anxiety disorders. For generalized anxiety, panic disorder, and social anxiety, they are the medications with the best combination of evidence and long-term safety we have.

Which ones actually get used

The SSRIs (sertraline, escitalopram, fluoxetine, paroxetine ) and the SNRIs (venlafaxine and duloxetine). Several carry specific FDA approvals across generalized anxiety, panic, and social anxiety disorder, though which drug is approved for which condition varies in ways that matter more to insurance companies than to biology.

In practice I choose based on things the approval label doesn't capture: what a patient has responded to before, what a first-degree relative responded to, how sensitive they are to sedation or activation, whether they're on anything with interaction potential, and how much a discontinuation problem would cost them if they miss doses.

How long before it works

Longer than most people are told, and that gap is where treatment fails.

Expect several weeks before meaningful change, and often longer for anxiety than for depression. Anxiety also tends to need higher doses than depression does to get the same benefit, which means the timeline includes not just waiting but titrating.

What I tell patients: the first thing you'll notice is side effects. The benefit shows up later, and it usually shows up as an absence rather than an event. Less time spent circling a worry. Fewer nights lying awake replaying a conversation. A phone call you make without rehearsing it first. Most people don't notice the day it starts working; they notice a few weeks in that something has gotten quieter.

The first two weeks are the hardest

This is the part I spend the most time on, because it's where people quit.

Serotonergic medications can make anxiety temporarily worse when you start them. More jittery, more keyed up, sometimes worse sleep. For someone who came in specifically because they feel wired and on edge, being handed a pill that makes them feel more wired and on edge is a compelling argument that the doctor was wrong.

It usually settles within a couple of weeks. Starting at a low dose and moving up slowly makes it much less likely to happen at all. But nobody rides that out unless they know it's coming, so I'd rather over-explain it than have someone stop on day four and conclude that medication doesn't work for them.

What about benzodiazepines?

They work, they work fast, and that's both the appeal and the problem.

A benzodiazepine will reduce acute anxiety within an hour. Nothing else in psychiatry does that. There are situations where that's genuinely the right tool: a short bridge while an SSRI takes effect, a specific and time-limited situation, occasional use for someone who has demonstrated they can use them occasionally.

The trouble is what happens with regular use. Tolerance builds, so the same dose does less. Stopping after sustained use is medically serious, not just uncomfortable. There are cognitive effects and, in older adults, fall risk. Combined with opioids or alcohol, the risk is respiratory. And there's a subtler problem: relief that arrives in twenty minutes teaches the brain that anxiety is an emergency requiring rescue, which works against the thing therapy is trying to build.

I prescribe them. I prescribe them carefully, usually with a defined role and an end point, and I'd rather have that conversation at the start than during a taper.

Other options worth knowing about

Buspirone treats generalized anxiety, has no dependence liability, and takes weeks to work. Modest but real, and useful as an add-on when an SSRI gets someone partway.

Hydroxyzine is an antihistamine that can be used as needed. Sedating, not habit-forming, better than a benzodiazepine for occasional use in someone where dependence is a concern.

Propranolol blocks the physical layer — racing heart, shaking hands, unsteady voice. Genuinely useful for performance situations. It does nothing for the worry itself, so it's a targeted tool rather than a treatment for an anxiety disorder.

The side effects worth remembering

Sexual side effects are common and underdiscussed. Reduced desire, delayed or absent orgasm, difficulty with arousal. Patients often won't raise it, and clinicians often don't ask, so people quietly stop taking a medication that was working and give a different reason. I ask directly, because there are usually options: dose adjustment, switching, or adding something.

Stopping matters as much as starting. Medications with shorter half-lives, paroxetine and venlafaxine especially, can produce a rough discontinuation syndrome if stopped abruptly or if doses get missed. This is not addiction and it doesn't mean the medication was dangerous, but it does mean these come off on a taper, not all at once.

Weight and sleep shift for some people, in both directions. Worth tracking rather than discovering months later.

Medication is one piece

Cognitive behavioral therapy, particularly with an exposure component, performs at least as well as medication for most anxiety disorders. It also has an advantage medication doesn't: the benefit tends to persist after treatment ends, because what you've built is a skill rather than a drug level.

For many people the combination beats either alone. For some, therapy alone is the right answer and medication never needs to enter the picture.

And then there's the part that sounds too simple to matter. Caffeine intake. Alcohol, which reliably worsens anxiety over the following day even when it helps in the moment. Sleep. Whether someone's life has any recovery time in it at all. An SSRI cannot fix a schedule with no margin in it. It can give someone enough room to change the schedule.

When it isn't working

Before concluding a medication failed, I want to know whether it actually got a fair trial: adequate dose, adequate duration, taken consistently. A meaningful share of "treatment-resistant" anxiety turns out to be an underdosed or too-brief trial.

If it did get a fair trial, the next move might be a dose increase, a switch within the class, a switch to a different class, or adding something. It's also worth revisiting the diagnosis. Anxiety that isn't responding as expected is sometimes anxiety that's actually something else, or anxiety sitting on top of a sleep disorder, a thyroid problem, or ongoing substance use.

The research supports antidepressants as a frontline option for most anxiety disorders. It doesn't tell me which one is right for you, at what dose, alongside what else, or whether you need medication at all. That's a conversation.

This is general information, not personalized medical advice. Don't start, stop, or change a medication without talking to your own clinician.