Ketamine Got Respectable. It Didn't Get Safe.

Matthew Perry died on October 28, 2023, unresponsive in the hot tub of his Los Angeles home, with ketamine from an infusion three and a half days earlier still acting on his heart. He wasn’t buying it off a dealer in a parking lot. He was getting it from doctors — one of whom texted a colleague calling Perry a “moron” who could be milked for cash, according to the Department of Justice’s account of the case. Dr. Salvador Plasencia was sentenced to 30 months in federal prison in December 2025. A second physician, Dr. Mark Chavez, got eight months of home confinement for supplying the ketamine that made it to Perry in the first place. Perry was in long-term recovery from opioid and alcohol addiction. The drug that killed him came with a prescription pad attached.

That’s the whole story of ketamine right now, compressed into one very famous death. This is the drug with a legitimate FDA-approved medical role, a booming national clinic industry treating depression, and a recreational culture — “Special K,” the K-hole, the club scene — that never went away and is, by most measures, growing. Those are not two different ketamines. It’s the same molecule, and right now the system deciding who administers it, how much, and to whom is mostly the honor system.

If you’ve been in recovery for any length of time, you already know how that sentence should make you feel.

The Clinic Boom Nobody Voted On

In 2015, there were fewer than 100 ketamine clinics operating in the United States. By 2024, that number had crossed 1,500, according to a January 2026 regulatory analysis in PharmExec — and industry trackers now put the figure north of 2,500 locations across all 50 states. Almost none of that growth is the one thing most people assume it is: the FDA-approved drug.

The only FDA-approved ketamine-class medication for psychiatric use is esketamine — brand name Spravato, the mirror-image “S” version of the ketamine molecule, delivered as a nasal spray under direct medical supervision. It’s approved for treatment-resistant depression (people who’ve failed at least two other antidepressants) and, combined with an oral antidepressant, for depression with active suicidal thoughts. In January 2025, the FDA expanded that approval so Spravato can be used alone, without a second antidepressant. That’s a real, closely regulated drug with a REMS program — a risk-management protocol that requires patients to be monitored on-site for at least two hours after each dose, precisely because dissociation and blood-pressure spikes are real risks.

Almost everything else you’ll find at a storefront “ketamine clinic” is off-label use of plain generic ketamine — an anesthetic approved in 1970 for surgical sedation, never for depression — given as an IV infusion, a lozenge, or, increasingly, a take-home nasal spray prescribed over a video call. The FDA has warned twice, in 2022 and again in October 2023, that compounded ketamine products are not FDA-approved for any psychiatric use and that using them without a provider physically present to monitor sedation, dissociation, and vital signs “may put patients at risk for serious adverse events.” As recently as June 2026, the agency was still sending warning letters to online sellers of unapproved, unmonitored ketamine products. The warnings keep coming because the market keeps outrunning them.

What the DEA Found When It Started Looking

For most of the clinic boom, oversight was mostly reactive. That’s shifting. The DEA has moved from chasing individual diversion cases to auditing clinic-level record-keeping, billing, and prescribing patterns — and it doesn’t need to prove anyone meant to break the rules. A sloppy inventory log is enough to put a registration at risk. Federal prosecutors have already charged clinic physicians in St. Louis with distribution conspiracy and healthcare fraud tied to minimal patient evaluations and unsupervised dosing, and the DEA suspended a South Carolina telemedicine prescriber’s registration over what it called large-scale prescribing without adequate examinations.

A sloppy inventory log is enough to put a registration at risk.

Layer telehealth on top of that and the picture gets messier. Pandemic-era rules that let providers prescribe controlled substances over video without an in-person exam have been repeatedly extended rather than resolved — the DEA’s latest extension runs through the end of 2026 — which means the at-home, mail-order side of the ketamine industry has been operating for years on a “temporary” flexibility nobody has committed to making permanent or ending. Plasencia and Chavez weren’t a telehealth case; they were two doctors with an in-person medical practice who decided a Hollywood addict’s money was worth the risk. The system failed in the room, not the inbox. That should worry you more, not less — it means better paperwork alone won’t fix this.

Angel Dust Already Ran This Experiment

Ketamine’s cousin already lived this arc, and it ended badly. PCP — phencyclidine — was developed in the 1950s as a surgical anesthetic, sold as Sernyl, used in operating rooms by 1963. Within a few years, patients were waking up agitated and psychotic, and PCP got pulled from human medicine by the late ’60s, demoted to a veterinary tranquilizer. Instead of disappearing, it became “angel dust” — a cheap ’70s and ’80s street drug tied to real psychotic breaks and enough moral panic that it landed in the DEA’s most restrictive schedule, manufacturing banned by 1978. Ketamine is retracing that same path, slower and better-marketed: battlefield anesthetic, then ’90s rave-scene K-hole, now wellness clinics. The difference is ketamine never fully lost its medical foothold — which is exactly why this round is harder to see clearly, not easier.

If Someone Offers You Ketamine Therapy in Recovery

Here’s where this stops being abstract. A 2025 analysis of substance use treatment populations found ketamine use markedly more common among people currently in treatment for a substance use disorder than among those who weren’t — the exact opposite of what you’d hope, and a real hint that dissociative use and recovery populations already overlap more than most treatment programs account for. NIDA is actively funding trials testing ketamine-assisted therapy for cocaine and methamphetamine addiction, and testing subanesthetic ketamine doses in emergency rooms for people with opioid use disorder. That research is legitimate and worth watching. It also means the referral pattern Perry lived — someone with a substance use history being offered a dissociative drug as treatment — isn’t a freak occurrence. It’s becoming a standard clinical conversation.

So say you’re two years off fentanyl, stable on your program, and your therapist raises ketamine-assisted therapy for depression that hasn’t moved. That’s not a trap and it’s not automatically a risk — Spravato under REMS supervision is a different animal than a storefront IV clinic that skips your intake history. But you’re allowed to ask hard questions before you say yes: Is this FDA-approved esketamine under a REMS protocol, or off-label generic ketamine? Will a licensed provider be physically in the room during dosing, not just watching on a screen? Does the clinic ask about your substance use history at all, or just your credit card?

If you’re the case manager making that referral: ask the clinic, in writing, before you send anyone — does intake include SUD history and current medications, is a prescriber physically on-site during administration, and what’s their protocol if a patient requests extra off-schedule doses. If they can’t answer plainly, don’t send your client there this week or any other week.

The honest limit here: nobody can tell you how often ketamine-clinic harm actually happens, because there’s no national registry tracking outcomes or adverse events across those 1,500-plus clinics. The safety data the industry cites is mostly self-reported by providers with a financial stake in the answer. That gap is the story as much as any single number in it.

The safety data the industry cites is mostly self-reported by providers with a financial stake in the answer.

Perry had years of hard-won recovery, resources most patients never get near, and a security system of doctors around him — and the system still handed him the drug that killed him. The clinic on the corner promising you fast relief from depression might be running a tight, careful, medically serious program. It might also be a storefront with a controlled-substance license and a waiting list. From the sidewalk, they look identical. That’s not a coincidence anyone’s fixing soon — it’s the business model.

For more on how dissociatives intersect with recovery, and what evidence-based treatment and recovery actually looks like, Rize’s newsroom keeps tracking where the regulation lags the reality.