I prescribe Xanax rarely, and when I do I usually regret it within six months. That’s not a slogan, it’s a clinical observation from watching how this drug plays out over time, and the math is the same just about every time. The drug works exactly the way it says on the tin… it shuts down anxiety in twenty minutes, which is also what makes it a problem. The thing that makes Xanax effective is the thing that makes it addictive, and pretending otherwise is what got us into the mess we’re currently in.

It’s a benzodiazepine, which means it sits on the GABA-A receptor and turns up the volume on your brain’s main “calm down” neurotransmitter, which quiets everything in a hurry. Xanax in particular has a short half-life, six to twelve hours, which is the worst possible combination of features for an anxiolytic (anxiety medication) if your goal is daily use… fast on, fast off, which means fast tolerance and a rebound anxiety profile on the way out that’s worse than the anxiety you started with.

The narrow defensible niche

Acute panic attacks where somebody’s going to the ER otherwise. Pre-procedure anxiety. A few weeks of bridge therapy while an SSRI is ramping up in somebody who can’t function in the meantime. That’s about it.

What it’s not: a daily medication for generalized anxiety, social anxiety, work stress, or sleep. It will work for those things in the short run, it will stop working in three to six months, and the off-ramp is brutal because benzo withdrawal is one of the few withdrawal syndromes that can actually kill you. Most withdrawal is just miserable. Benzo withdrawal includes seizures. The math on starting a chronic benzo for chronic anxiety almost never pencils out once you look five years down the road.

The patient-autonomy piece, because it’s awkward here

If somebody walks in and wants a benzo and the case is reasonable, I’m a provider, not a parent. My job is the honest take, theirs is the choice. The honest take with Xanax is that it works fast in the short run and it builds itself a problem in the long run, and most people who tell me their last doctor had them on it for years are walking around in a worse anxiety state than they would have been on an SSRI with CBT. With Xanax specifically, my “disapproving yes” rate is the highest in my prescribing pattern, meaning I’ll write the short-term script when somebody asks and is in the right situation for it, with a clear off-ramp planned from day one, and I’ll document the conversation honestly. I hardly ever say no. I do say “let’s talk about what this looks like in eighteen months” a lot.

If what you’re shopping for is a prescriber who’ll keep refilling this for years and stop asking why, that’s not me. I’m not a no, I’m a yes-with-questions, and the questions aren’t going anywhere. The script comes with the conversation, every visit, for as long as the prescription does.

What guys come in saying

“My last doctor put me on 1 mg three times a day and it works great.” Yeah, I bet it does. And in eighteen months when you try to come off it because the rest of your life has gotten complicated, you’re going to learn that benzo withdrawal includes seizures, that you can’t just stop, that the taper takes nine to fifteen months done correctly, and that the version of anxiety you have on the way off is going to be a different and worse animal than the anxiety you originally walked in with. None of that is in the patient education leaflet. The leaflet’s job is to make you feel safe enough to start the drug.

The other version: “I only take it when I need it.” Sometimes true. Sometimes the definition of “need it” slowly broadens. Sometimes the brain learns that any uncomfortable feeling has a fix in the medicine cabinet, and the brain stops doing the work of tolerating discomfort, which is most of what anxiety treatment actually is. You can put lipstick on a pig but it’s still a pig.

What to try first instead

An SSRI like Lexapro, sertraline, or paroxetine. First-line for generalized anxiety, social anxiety, and panic disorder, with actual long-term data behind them. The trade-off (some sexual side effects, six weeks of ramp-up) is small compared to the trade-off of being on a benzo in three years and trying to taper off.

Buspirone is the next stop… it’s a different kind of anti-anxiety drug, hits a specific serotonin receptor without going anywhere near the benzo wiring, no addiction potential, modestly effective for generalized anxiety. Not a heavy hitter, not for panic, but a real option for the daily-low-grade-worry guy. Underused.

Hydroxyzine is an antihistamine (yes, the same category as Benadryl, sort of) that works as a PRN anxiolytic (PRN means “as needed,” meaning you take it when you actually need it, not on a schedule). Less effective than a benzo, also less addictive (read: not addictive at all), causes drowsiness which can be a feature or a bug depending on context. I write this a lot for situational anxiety where somebody needs something in their pocket.

Propranolol is a beta-blocker, originally a blood pressure drug, that doesn’t touch the anxious feeling itself but kills the physical symptoms (racing heart, shaking, sweating)… excellent for performance anxiety, public speaking, social situations, and the bonus is it doesn’t cause dependence and it’s cheap.

And yes, CBT (cognitive behavioral therapy, the structured worksheet-and-homework kind, see the therapy posts for the longer conversation), specifically exposure-based CBT for panic disorder has data that rivals medication, and the effects last after you stop the work, which medication effects do not. If you can do the work, do the work.

If we’re going to use it, here’s how

Lowest effective dose. 0.25 mg or 0.5 mg, not 1 mg. Use it PRN (as-needed) and not on a schedule, limit to a week or two if at all possible, tell the patient up front before the first dose that this isn’t a long-term plan and what the off-ramp will look like if it becomes one, document the conversation honestly, refill carefully, and don’t pretend the conversation didn’t happen at the six-month mark when the patient wants to extend.

And honestly… half the time when I do write a small supply, it sits in the medicine cabinet for two years and never gets used, because just knowing it’s there for the worst-case panic attack is enough to settle the patient down. Which is its own interesting thing about anxiety and which is also part of why the medication’s role in this is smaller than the prescribing rate would suggest.

What’s nice to hear about the alternatives

The thing nobody opens with, because benzo content has gotten so heavy on the alarm side, is that the alternatives actually work for most guys. SSRIs at a real dose plus CBT plus a propranolol or hydroxyzine in the pocket for the worst situations is a combination that gets most people from “I can’t function” to “I can function” inside three months, and at the eighteen-month mark they don’t have a dependence problem to taper out of. The kind of guy who finishes a thoughtful version of the SSRI-plus-therapy track usually ends up in better shape than the guy who got handed a Xanax script on his second appointment, and the difference shows up most clearly at the three-year mark when one of them is white-knuckling a taper and the other one isn’t. The non-benzo route is a real, functional, often boring success story that doesn’t get written about because it doesn’t have the drama.

The pattern that ends up tapering

The kind of guy who comes in already on chronic Xanax is usually somebody whose primary care doc has been refilling it for six years after a life event that was supposed to be the temporary stressor. He doesn’t think he has a problem with it, his wife does, he says it’s the only thing that’s worked. The right question to ask is how long he’s been telling himself it’s the only thing that’s worked.

We do the taper. Switch to Klonopin first, which is the longer-half-life benzo and is genuinely the better tool for stepping down, then drop the dose by about ten percent a month for eleven months. Start sertraline at month three once the worst of the rebound anxiety is past. Add CBT. There are usually a couple of rough patches around months four and seven where I think we’re going to have to slow down further, but we usually get through them. Two years later he’s off benzos, on 100 mg of sertraline, and he tells me at the visit that the version of himself who needed Xanax every day was a guy who’d stopped believing he could handle his own life. Now he handles his own life, and sometimes it sucks, and he handles it anyway, which is most of what getting better looks like for guys like him.

Note on Klonopin specifically, because there’s a temptation to think it’s the “cleaner benzo.” It isn’t, it’s a slower benzo… which sounds like it should be safer but isn’t really, the longer half-life just means the same problem unfolds more slowly. The one real argument for Klonopin in this whole context is as a taper tool to step somebody down off a shorter-acting benzo like Xanax, exactly because the slower clearance smooths out the dose drops. As a chronic daily medication it has the same long-run shape as Xanax does, just stretched out.

The thing that makes Xanax effective is the thing that makes it addictive.

What not to do

Don’t combine with opioids. The respiratory depression stack is what kills people, and the bulk of benzo-related overdose deaths have an opioid co-involved.

Don’t drink on it, same reason, same math.

Don’t stop cold turkey if you’ve been on it more than a month or two. Get a taper plan. Withdrawal seizures are real and they happen to people who didn’t think they were dependent.

Don’t buy them off your friend, off the internet, or off anyone whose business cards aren’t laminated. The counterfeit alprazolam supply is laced with fentanyl now, in some cases pressed into pills that look identical to the real thing. People are dying from this. Not to be Chicken Little about it, but: this is the actual road that’s killing people in the news this year, not the version where you took a real prescription Xanax. If you can’t get a prescription, the answer isn’t the friend with a stash, the answer is a different prescriber.

If you’ve been on it a while and want to come off

The taper is the project. The Ashton Manual is the patient-facing classic and it isn’t wrong… the basic idea is to switch to a longer-half-life benzo (usually Klonopin or Valium) for the taper itself, because the longer half-life makes the dose drops less obvious to the body, and then come down by about five to ten percent of the current dose every two to four weeks. Some patients need to go slower toward the end, where each proportional drop feels bigger. The whole project is usually six to fifteen months done right, which sounds like a lot until you compare it to the alternative of cold-turkey withdrawal that can include seizures.

Starting an SSRI three to six weeks into the taper, before the worst of the rebound anxiety hits, is the usual move. Adding CBT is the other usual move. Both are part of why guys who do the taper well end up in better shape than they were on the benzo in the first place. The drug was treating the surface. The real work of building underlying anxiety tolerance was getting skipped because the drug was making it unnecessary in the short run, and once you actually do the work, your future-self ends up with a wider operating range than the medicated version of you ever had.

Bottom line

Benzos in general and Xanax in particular have a narrow place in modern psychiatry, and the place they’ve actually occupied is not that narrow place. If you’re already on it daily and have been for a while, we’re not going to yank it out from under you, we’re going to put together a real taper. If you’re being offered it as a daily medication for ongoing anxiety, ask about literally any of the alternatives first. The decision is yours either way. The honest take is the one above.

Sources

U.S. Food and Drug Administration. Xanax (alprazolam) Prescribing Information. NDA 018276. FDA; 2023. FDA label.
Du Y, Du B, Diao Y, et al. Comparative efficacy and acceptability of antidepressants and benzodiazepines for the treatment of panic disorder: A systematic review and network meta-analysis. Asian J Psychiatr. 2021;60:102664. PMID 33965693.
Slee A, Nazareth I, Bondaronek P, et al. Pharmacological treatments for generalised anxiety disorder: a systematic review and network meta-analysis. Lancet. 2019;393(10173):768–777. PMID 30712879.